bump version number for release 2.3.5

# Distinguish LoRA/LyCORIS files based on what version of SD they were
built on top of

- Attempting to run a prompt with a LoRA based on SD v1.X against a
model based on v2.X will now throw an `IncompatibleModelException`. To
import this exception:
`from ldm.modules.lora_manager import IncompatibleModelException` (maybe
this should be defined in ModelManager?)
    
- Enhance `LoraManager.list_loras()` to accept an optional integer
argument, `token_vector_length`. This will filter the returned LoRA
models to return only those that match the indicated length. Use:
      ```
      768 => for models based on SD v1.X
      1024 => for models based on SD v2.X
      ```
Note that this filtering requires each LoRA file to be opened by
`torch.safetensors`. It will take ~8s to scan a directory of 40 files.
    
- Added new static methods to `ldm.modules.kohya_lora_manager`:
      - check_model_compatibility()
      - vector_length_from_checkpoint()
      - vector_length_from_checkpoint_file()

- You can now create subdirectories within the `loras` directory and
organize the model files.

.coveragerc

@@ -0,0 +1,6 @@
+[run]
+omit='.env/*'
+source='.'
+
+[report]
+show_missing = true

.github/CODEOWNERS

@@ -1,13 +1,13 @@
 # continuous integration
-/.github/workflows/ @mauwii @lstein @blessedcoolant
+/.github/workflows/  @lstein @blessedcoolant
 
 # documentation
-/docs/ @lstein @mauwii @blessedcoolant
-mkdocs.yml @mauwii @lstein
+/docs/ @lstein  @blessedcoolant
+mkdocs.yml  @lstein @ebr
 
 # installation and configuration
-/pyproject.toml @mauwii @lstein @ebr
-/docker/ @mauwii
+/pyproject.toml  @lstein @ebr
+/docker/ @lstein
 /scripts/ @ebr @lstein @blessedcoolant
 /installer/ @ebr @lstein 
 ldm/invoke/config @lstein @ebr
@@ -21,13 +21,13 @@ invokeai/configs @lstein @ebr  @blessedcoolant
 
 # generation and model management
 /ldm/*.py @lstein @blessedcoolant
-/ldm/generate.py @lstein @keturn
+/ldm/generate.py @lstein @gregghelt2
 /ldm/invoke/args.py @lstein @blessedcoolant
 /ldm/invoke/ckpt* @lstein @blessedcoolant
 /ldm/invoke/ckpt_generator @lstein @blessedcoolant
 /ldm/invoke/CLI.py @lstein @blessedcoolant
-/ldm/invoke/config @lstein @ebr @mauwii @blessedcoolant
-/ldm/invoke/generator @keturn @damian0815
+/ldm/invoke/config @lstein @ebr  @blessedcoolant
+/ldm/invoke/generator @gregghelt2 @damian0815
 /ldm/invoke/globals.py @lstein @blessedcoolant 
 /ldm/invoke/merge_diffusers.py @lstein @blessedcoolant
 /ldm/invoke/model_manager.py @lstein @blessedcoolant
@@ -36,17 +36,17 @@ invokeai/configs @lstein @ebr  @blessedcoolant
 /ldm/invoke/restoration @lstein @blessedcoolant
 
 # attention, textual inversion, model configuration
-/ldm/models @damian0815 @keturn @blessedcoolant
+/ldm/models @damian0815 @gregghelt2 @blessedcoolant
 /ldm/modules/textual_inversion_manager.py @lstein @blessedcoolant
-/ldm/modules/attention.py @damian0815 @keturn
-/ldm/modules/diffusionmodules @damian0815 @keturn
-/ldm/modules/distributions @damian0815 @keturn
-/ldm/modules/ema.py  @damian0815 @keturn
+/ldm/modules/attention.py @damian0815 @gregghelt2
+/ldm/modules/diffusionmodules @damian0815 @gregghelt2
+/ldm/modules/distributions @damian0815 @gregghelt2
+/ldm/modules/ema.py  @damian0815 @gregghelt2
 /ldm/modules/embedding_manager.py @lstein
-/ldm/modules/encoders @damian0815 @keturn
-/ldm/modules/image_degradation @damian0815 @keturn
-/ldm/modules/losses  @damian0815 @keturn
-/ldm/modules/x_transformer.py @damian0815 @keturn
+/ldm/modules/encoders @damian0815 @gregghelt2
+/ldm/modules/image_degradation @damian0815 @gregghelt2
+/ldm/modules/losses  @damian0815 @gregghelt2
+/ldm/modules/x_transformer.py @damian0815 @gregghelt2
 
 # Nodes
 apps/ @Kyle0654 @jpphoto

.gitignore

@@ -68,6 +68,7 @@ htmlcov/
 .cache
 nosetests.xml
 coverage.xml
+cov.xml
 *.cover
 *.py,cover
 .hypothesis/

.pytest.ini

@@ -0,0 +1,5 @@
+[pytest]
+DJANGO_SETTINGS_MODULE = webtas.settings
+; python_files = tests.py test_*.py *_tests.py
+
+addopts = --cov=. --cov-config=.coveragerc --cov-report xml:cov.xml

docs/contributing/ARCHITECTURE.md

@@ -0,0 +1,93 @@
+# Invoke.AI Architecture
+
+```mermaid
+flowchart TB
+
+  subgraph apps[Applications]
+    webui[WebUI]
+    cli[CLI]
+
+  subgraph webapi[Web API]
+    api[HTTP API]
+    sio[Socket.IO]
+  end
+
+  end
+
+  subgraph invoke[Invoke]
+    direction LR
+    invoker
+    services
+    sessions
+    invocations
+  end
+
+  subgraph core[AI Core]
+    Generate
+  end
+
+  webui --> webapi
+  webapi --> invoke
+  cli --> invoke
+
+  invoker --> services & sessions
+  invocations --> services
+  sessions --> invocations
+
+  services --> core
+
+  %% Styles
+  classDef sg fill:#5028C8,font-weight:bold,stroke-width:2,color:#fff,stroke:#14141A
+  classDef default stroke-width:2px,stroke:#F6B314,color:#fff,fill:#14141A
+
+  class apps,webapi,invoke,core sg
+
+```
+
+## Applications
+
+Applications are built on top of the invoke framework. They should construct `invoker` and then interact through it. They should avoid interacting directly with core code in order to support a variety of configurations.
+
+### Web UI
+
+The Web UI is built on top of an HTTP API built with [FastAPI](https://fastapi.tiangolo.com/) and [Socket.IO](https://socket.io/). The frontend code is found in `/frontend` and the backend code is found in `/ldm/invoke/app/api_app.py` and `/ldm/invoke/app/api/`. The code is further organized as such:
+
+| Component | Description |
+| --- | --- |
+| api_app.py | Sets up the API app, annotates the OpenAPI spec with additional data, and runs the API |
+| dependencies | Creates all invoker services and the invoker, and provides them to the API |
+| events | An eventing system that could in the future be adapted to support horizontal scale-out |
+| sockets | The Socket.IO interface - handles listening to and emitting session events (events are defined in the events service module) |
+| routers | API definitions for different areas of API functionality |
+
+### CLI
+
+The CLI is built automatically from invocation metadata, and also supports invocation piping and auto-linking. Code is available in `/ldm/invoke/app/cli_app.py`.
+
+## Invoke
+
+The Invoke framework provides the interface to the underlying AI systems and is built with flexibility and extensibility in mind. There are four major concepts: invoker, sessions, invocations, and services.
+
+### Invoker
+
+The invoker (`/ldm/invoke/app/services/invoker.py`) is the primary interface through which applications interact with the framework. Its primary purpose is to create, manage, and invoke sessions. It also maintains two sets of services:
+- **invocation services**, which are used by invocations to interact with core functionality.
+- **invoker services**, which are used by the invoker to manage sessions and manage the invocation queue.
+
+### Sessions
+
+Invocations and links between them form a graph, which is maintained in a session. Sessions can be queued for invocation, which will execute their graph (either the next ready invocation, or all invocations). Sessions also maintain execution history for the graph (including storage of any outputs). An invocation may be added to a session at any time, and there is capability to add and entire graph at once, as well as to automatically link new invocations to previous invocations. Invocations can not be deleted or modified once added.
+
+The session graph does not support looping. This is left as an application problem to prevent additional complexity in the graph.
+
+### Invocations
+
+Invocations represent individual units of execution, with inputs and outputs. All invocations are located in `/ldm/invoke/app/invocations`, and are all automatically discovered and made available in the applications. These are the primary way to expose new functionality in Invoke.AI, and the [implementation guide](INVOCATIONS.md) explains how to add new invocations.
+
+### Services
+
+Services provide invocations access AI Core functionality and other necessary functionality (e.g. image storage). These are available in `/ldm/invoke/app/services`. As a general rule, new services should provide an interface as an abstract base class, and may provide a lightweight local implementation by default in their module. The goal for all services should be to enable the usage of different implementations (e.g. using cloud storage for image storage), but should not load any module dependencies unless that implementation has been used (i.e. don't import anything that won't be used, especially if it's expensive to import).
+
+## AI Core
+
+The AI Core is represented by the rest of the code base (i.e. the code outside of `/ldm/invoke/app/`).

docs/contributing/INVOCATIONS.md

@@ -0,0 +1,105 @@
+# Invocations
+
+Invocations represent a single operation, its inputs, and its outputs. These operations and their outputs can be chained together to generate and modify images.
+
+## Creating a new invocation
+
+To create a new invocation, either find the appropriate module file in `/ldm/invoke/app/invocations` to add your invocation to, or create a new one in that folder. All invocations in that folder will be discovered and made available to the CLI and API automatically. Invocations make use of [typing](https://docs.python.org/3/library/typing.html) and [pydantic](https://pydantic-docs.helpmanual.io/) for validation and integration into the CLI and API.
+
+An invocation looks like this:
+
+```py
+class UpscaleInvocation(BaseInvocation):
+    """Upscales an image."""
+    type: Literal['upscale'] = 'upscale'
+
+    # Inputs
+    image: Union[ImageField,None] = Field(description="The input image")
+    strength: float               = Field(default=0.75, gt=0, le=1, description="The strength")
+    level: Literal[2,4]           = Field(default=2, description = "The upscale level")
+
+    def invoke(self, context: InvocationContext) -> ImageOutput:
+        image = context.services.images.get(self.image.image_type, self.image.image_name)
+        results = context.services.generate.upscale_and_reconstruct(
+            image_list     = [[image, 0]],
+            upscale        = (self.level, self.strength),
+            strength       = 0.0, # GFPGAN strength
+            save_original  = False,
+            image_callback = None,
+        )
+
+        # Results are image and seed, unwrap for now
+        # TODO: can this return multiple results?
+        image_type = ImageType.RESULT
+        image_name = context.services.images.create_name(context.graph_execution_state_id, self.id)
+        context.services.images.save(image_type, image_name, results[0][0])
+        return ImageOutput(
+            image = ImageField(image_type = image_type, image_name = image_name)
+        )
+```
+
+Each portion is important to implement correctly.
+
+### Class definition and type
+```py
+class UpscaleInvocation(BaseInvocation):
+    """Upscales an image."""
+    type: Literal['upscale'] = 'upscale'
+```
+All invocations must derive from `BaseInvocation`. They should have a docstring that declares what they do in a single, short line. They should also have a `type` with a type hint that's `Literal["command_name"]`, where `command_name` is what the user will type on the CLI or use in the API to create this invocation. The `command_name` must be unique. The `type` must be assigned to the value of the literal in the type hint.
+
+### Inputs
+```py
+    # Inputs
+    image: Union[ImageField,None] = Field(description="The input image")
+    strength: float               = Field(default=0.75, gt=0, le=1, description="The strength")
+    level: Literal[2,4]           = Field(default=2, description="The upscale level")
+```
+Inputs consist of three parts: a name, a type hint, and a `Field` with default, description, and validation information. For example:
+| Part | Value | Description |
+| ---- | ----- | ----------- |
+| Name | `strength` | This field is referred to as `strength` |
+| Type Hint | `float` | This field must be of type `float` |
+| Field | `Field(default=0.75, gt=0, le=1, description="The strength")` | The default value is `0.75`, the value must be in the range (0,1], and help text will show "The strength" for this field. |
+
+Notice that `image` has type `Union[ImageField,None]`. The `Union` allows this field to be parsed with `None` as a value, which enables linking to previous invocations. All fields should either provide a default value or allow `None` as a value, so that they can be overwritten with a linked output from another invocation.
+
+The special type `ImageField` is also used here. All images are passed as `ImageField`, which protects them from pydantic validation errors (since images only ever come from links).
+
+Finally, note that for all linking, the `type` of the linked fields must match. If the `name` also matches, then the field can be **automatically linked** to a previous invocation by name and matching.
+
+### Invoke Function
+```py
+    def invoke(self, context: InvocationContext) -> ImageOutput:
+        image = context.services.images.get(self.image.image_type, self.image.image_name)
+        results = context.services.generate.upscale_and_reconstruct(
+            image_list     = [[image, 0]],
+            upscale        = (self.level, self.strength),
+            strength       = 0.0, # GFPGAN strength
+            save_original  = False,
+            image_callback = None,
+        )
+
+        # Results are image and seed, unwrap for now
+        image_type = ImageType.RESULT
+        image_name = context.services.images.create_name(context.graph_execution_state_id, self.id)
+        context.services.images.save(image_type, image_name, results[0][0])
+        return ImageOutput(
+            image = ImageField(image_type = image_type, image_name = image_name)
+        )
+```
+The `invoke` function is the last portion of an invocation. It is provided an `InvocationContext` which contains services to perform work as well as a `session_id` for use as needed. It should return a class with output values that derives from `BaseInvocationOutput`.
+
+Before being called, the invocation will have all of its fields set from defaults, inputs, and finally links (overriding in that order).
+
+Assume that this invocation may be running simultaneously with other invocations, may be running on another machine, or in other interesting scenarios. If you need functionality, please provide it as a service in the `InvocationServices` class, and make sure it can be overridden.
+
+### Outputs
+```py
+class ImageOutput(BaseInvocationOutput):
+    """Base class for invocations that output an image"""
+    type: Literal['image'] = 'image'
+
+    image: ImageField = Field(default=None, description="The output image")
+```
+Output classes look like an invocation class without the invoke method. Prefer to use an existing output class if available, and prefer to name inputs the same as outputs when possible, to promote automatic invocation linking.

docs/features/CONCEPTS.md

@@ -1,5 +1,5 @@
 ---
-title: Concepts Library
+title: Styles and Subjects
 ---
 
 # :material-library-shelves: The Hugging Face Concepts Library and Importing Textual Inversion files
@@ -25,10 +25,14 @@ library which downloads and merges TI files automatically upon request. You can
 also install your own or others' TI files by placing them in a designated
 directory.
 
+You may also be interested in using [LoRA Models](LORAS.md) to
+generate images with specialized styles and subjects.
+
 ### An Example
 
-Here are a few examples to illustrate how it works. All these images were
-generated using the command-line client and the Stable Diffusion 1.5 model:
+Here are a few examples to illustrate how Textual Inversion works. All
+these images were generated using the command-line client and the
+Stable Diffusion 1.5 model:
 
 |         Japanese gardener          | Japanese gardener <ghibli-face> | Japanese gardener <hoi4-leaders> | Japanese gardener <cartoona-animals> |
 | :--------------------------------: | :-----------------------------------: | :------------------------------------: | :----------------------------------------: |
@@ -109,21 +113,50 @@ For example, TI files generated by the Hugging Face toolkit share the named
 `learned_embedding.bin`. You can use subdirectories to keep them distinct.
 
 At startup time, InvokeAI will scan the `embeddings` directory and load any TI
-files it finds there. At startup you will see a message similar to this one:
+files it finds there. At startup you will see messages similar to these:
 
 ```bash
->> Current embedding manager terms: *, <HOI4-Leader>, <princess-knight>
+>> Loading embeddings from /data/lstein/invokeai-2.3/embeddings
+   | Loading v1 embedding file: style-hamunaptra
+   | Loading v4 embedding file: embeddings/learned_embeds-steps-500.bin
+   | Loading v2 embedding file: lfa
+   | Loading v3 embedding file: easynegative
+   | Loading v1 embedding file: rem_rezero
+   | Loading v2 embedding file: midj-strong
+   | Loading v4 embedding file: anime-background-style-v2/learned_embeds.bin
+   | Loading v4 embedding file: kamon-style/learned_embeds.bin
+   ** Notice: kamon-style/learned_embeds.bin was trained on a model with an incompatible token dimension: 768 vs 1024.
+>> Textual inversion triggers: <anime-background-style-v2>, <easynegative>, <lfa>, <midj-strong>, <milo>, Rem3-2600, Style-Hamunaptra
 ```
 
-Note the `*` trigger term. This is a placeholder term that many early TI
-tutorials taught people to use rather than a more descriptive term.
-Unfortunately, if you have multiple TI files that all use this term, only the
-first one loaded will be triggered by use of the term.
+Textual Inversion embeddings trained on version 1.X stable diffusion
+models are incompatible with version 2.X models and vice-versa.
 
-To avoid this problem, you can use the `merge_embeddings.py` script to merge two
-or more TI files together. If it encounters a collision of terms, the script
-will prompt you to select new terms that do not collide. See
-[Textual Inversion](TEXTUAL_INVERSION.md) for details.
+After the embeddings load, InvokeAI will print out a list of all the
+recognized trigger terms. To trigger the term, include it in the
+prompt exactly as written, including angle brackets if any and
+respecting the capitalization.
+
+There are at least four different embedding file formats, and each uses
+a different convention for the trigger terms. In some cases, the
+trigger term is specified in the file contents and may or may not be
+surrounded by angle brackets. In the example above, `Rem3-2600`,
+`Style-Hamunaptra`, and `<midj-strong>` were specified this way and
+there is no easy way to change the term.
+
+In other cases the trigger term is not contained within the embedding
+file. In this case, InvokeAI constructs a trigger term consisting of
+the base name of the file (without the file extension) surrounded by
+angle brackets. In the example above `<easynegative`> is such a file
+(the filename was `easynegative.safetensors`). In such cases, you can
+change the trigger term simply by renaming the file.
+
+## Training your own Textual Inversion models
+
+InvokeAI provides a script that lets you train your own Textual
+Inversion embeddings using a small number (about a half-dozen) images
+of your desired style or subject. Please see [Textual
+Inversion](TEXTUAL_INVERSION.md) for details.
 
 ## Further Reading
 

docs/features/LORAS.md

@@ -0,0 +1,100 @@
+---
+title: Low-Rank Adaptation (LoRA) Models
+---
+
+# :material-library-shelves: Using Low-Rank Adaptation (LoRA) Models
+
+## Introduction
+
+LoRA is a technique for fine-tuning Stable Diffusion models using much
+less time and memory than traditional training techniques. The
+resulting model files are much smaller than full model files, and can
+be used to generate specialized styles and subjects.
+
+LoRAs are built on top of Stable Diffusion v1.x or 2.x checkpoint or
+diffusers models. To load a LoRA, you include its name in the text
+prompt using a simple syntax described below. While you will generally
+get the best results when you use the same model the LoRA was trained
+on, they will work to a greater or lesser extent with other models.
+The major caveat is that a LoRA built on top of a SD v1.x model cannot
+be used with a v2.x model, and vice-versa. If you try, you will get an
+error! You may refer to multiple LoRAs in your prompt.
+
+When you apply a LoRA in a prompt you can specify a weight. The higher
+the weight, the more influence it will have on the image. Useful
+ranges for weights are usually in the 0.0 to 1.0 range (with ranges
+between 0.5 and 1.0 being most typical). However you can specify a
+higher weight if you wish. Like models, each LoRA has a slightly
+different useful weight range and will interact with other generation
+parameters such as the CFG, step count and sampler. The author of the
+LoRA will often provide guidance on the best settings, but feel free
+to experiment. Be aware that it often helps to reduce the CFG value
+when using LoRAs.
+
+## Installing LoRAs
+
+This is very easy! Download a LoRA model file from your favorite site
+(e.g. [CIVITAI](https://civitai.com) and place it in the `loras`
+folder in the InvokeAI root directory (usually `~invokeai/loras` on
+Linux/Macintosh machines, and `C:\Users\your-name\invokeai/loras` on
+Windows systems). If the `loras` folder does not already exist, just
+create it. The vast majority of LoRA models use the Kohya file format,
+which is a type of `.safetensors` file.
+
+You may change where InvokeAI looks for the `loras` folder by passing the
+`--lora_directory` option to the `invoke.sh`/`invoke.bat` launcher, or
+by placing the option in `invokeai.init`. For example:
+
+```
+invoke.sh --lora_directory=C:\Users\your-name\SDModels\lora
+```
+
+## Using a LoRA in your prompt
+
+To activate a LoRA use the syntax `withLora(my-lora-name,weight)`
+somewhere in the text of the prompt. The position doesn't matter; use
+whatever is most comfortable for you.
+
+For example, if you have a LoRA named `parchment_people.safetensors`
+in your `loras` directory, you can load it with a weight of 0.9 with a
+prompt like this one:
+
+```
+family sitting at dinner table withLora(parchment_people,0.9)
+```
+
+Add additional `withLora()` phrases to load more LoRAs.
+
+You may omit the weight entirely to default to a weight of 1.0:
+
+```
+family sitting at dinner table withLora(parchment_people)
+```
+
+If you watch the console as your prompt executes, you will see
+messages relating to the loading and execution of the LoRA. If things
+don't work as expected, note down the console messages and report them
+on the InvokeAI Issues pages or Discord channel.
+
+That's pretty much all you need to know!
+
+## Training Kohya Models
+
+InvokeAI cannot currently train LoRA models, but it can load and use
+existing LoRA ones to generate images. While there are several LoRA
+model file formats, the predominant one is ["Kohya"
+format](https://github.com/kohya-ss/sd-scripts), written by [Kohya
+S.](https://github.com/kohya-ss). InvokeAI provides support for this
+format. For creating your own Kohya models, we recommend the Windows
+GUI written by former InvokeAI-team member
+[bmaltais](https://github.com/bmaltais), which can be found at
+[kohya_ss](https://github.com/bmaltais/kohya_ss).
+
+We can also recommend the [HuggingFace DreamBooth Training
+UI](https://huggingface.co/spaces/lora-library/LoRA-DreamBooth-Training-UI),
+a paid service that supports both Textual Inversion and LoRA training.
+
+You may also be interested in [Textual
+Inversion](TEXTUAL_INVERSION.md) training, which is supported by
+InvokeAI as a text console and command-line tool.
+

docs/features/index.md

@@ -20,6 +20,8 @@ title: Overview
 
         Scriptable access to InvokeAI's features.
 
+    -   [Visual Manual for InvokeAI](https://docs.google.com/presentation/d/e/2PACX-1vSE90aC7bVVg0d9KXVMhy-Wve-wModgPFp7AGVTOCgf4xE03SnV24mjdwldolfCr59D_35oheHe4Cow/pub?start=false&loop=true&delayms=60000) (contributed by Statcomm)
+
 -   Image Generation
 
     -   [Prompt Engineering](PROMPTS.md)

docs/index.md

@@ -142,6 +142,10 @@ This method is recommended for those familiar with running Docker containers
 -   [WebUI overview](features/WEB.md)
 -   [WebUI hotkey reference guide](features/WEBUIHOTKEYS.md)
 -   [WebUI Unified Canvas for Img2Img, inpainting and outpainting](features/UNIFIED_CANVAS.md)
+-   [Visual Manual for InvokeAI v2.3.1](https://docs.google.com/presentation/d/e/2PACX-1vSE90aC7bVVg0d9KXVMhy-Wve-wModgPFp7AGVTOCgf4xE03SnV24mjdwldolfCr59D_35oheHe4Cow/pub?start=false&loop=true&delayms=60000) (contributed by Statcomm)
+
+<!-- separator -->
+
 <!-- separator -->
 
 ### The InvokeAI Command Line Interface
@@ -155,6 +159,7 @@ This method is recommended for those familiar with running Docker containers
 -   [Inpainting](features/INPAINTING.md)
 -   [Outpainting](features/OUTPAINTING.md)
 -   [Adding custom styles and subjects](features/CONCEPTS.md)
+-   [Using LoRA models](features/LORAS.md)
 -   [Upscaling and Face Reconstruction](features/POSTPROCESS.md)
 -   [Embiggen upscaling](features/EMBIGGEN.md)
 -   [Other Features](features/OTHER.md)
@@ -165,7 +170,7 @@ This method is recommended for those familiar with running Docker containers
 
 -   [Installing](installation/050_INSTALLING_MODELS.md)
 -   [Model Merging](features/MODEL_MERGING.md)
--   [Style/Subject Concepts and Embeddings](features/CONCEPTS.md)
+-   [Adding custom styles and subjects via embeddings](features/CONCEPTS.md)
 -   [Textual Inversion](features/TEXTUAL_INVERSION.md)
 -   [Not Safe for Work (NSFW) Checker](features/NSFW.md)
 <!-- seperator -->
@@ -177,6 +182,154 @@ This method is recommended for those familiar with running Docker containers
 
 ## :octicons-log-16: Latest Changes
 
+### v2.3.3 <small>(29 March 2023)</small>
+
+#### Bug Fixes
+1. When using legacy checkpoints with an external VAE, the VAE file is now scanned for malware prior to loading. Previously only the main model weights file was scanned.
+2. Textual inversion will select an appropriate batchsize based on whether `xformers` is active, and will default to `xformers` enabled if the library is detected.
+3. The batch script log file names have been fixed to be compatible with Windows.
+4. Occasional corruption of the `.next_prefix` file (which stores the next output file name in sequence) on Windows systems is now detected and corrected.
+5. An infinite loop when opening the developer's console from within the `invoke.sh` script has been corrected.
+
+#### Enhancements
+1. It is now possible to load and run several community-contributed SD-2.0 based models, including the infamous "Illuminati" model.
+2. The "NegativePrompts" embedding file, and others like it, can now be loaded by placing it in the InvokeAI `embeddings` directory.
+3. If no `--model` is specified at launch time, InvokeAI will remember the last model used and restore it the next time it is launched.
+4. On Linux systems, the `invoke.sh` launcher now uses a prettier console-based interface. To take advantage of it, install the `dialog` package using your package manager (e.g. `sudo apt install dialog`).
+5. When loading legacy models (safetensors/ckpt) you can specify a custom config file and/or a VAE by placing like-named files in the same directory as the model following this example:
+```
+my-favorite-model.ckpt
+my-favorite-model.yaml
+my-favorite-model.vae.pt      # or my-favorite-model.vae.safetensors
+```
+
+### v2.3.2 <small>(13 March 2023)</small>
+
+#### Bugfixes
+
+Since version 2.3.1 the following bugs have been fixed:
+
+1. Black images appearing for potential NSFW images when generating with legacy checkpoint models and both `--no-nsfw_checker` and `--ckpt_convert` turned on.
+2. Black images appearing when generating from models fine-tuned on Stable-Diffusion-2-1-base. When importing V2-derived models, you may be asked to select whether the model was derived from a "base" model (512 pixels) or the 768-pixel SD-2.1 model.
+3. The "Use All" button was not restoring the Hi-Res Fix setting on the WebUI
+4. When using the model installer console app, models failed to import correctly when importing from directories with spaces in their names. A similar issue with the output directory was also fixed.
+5. Crashes that occurred during model merging.
+6. Restore previous naming of Stable Diffusion base and 768 models.
+7. Upgraded to latest versions of `diffusers`, `transformers`, `safetensors` and `accelerate` libraries upstream. We hope that this will fix the `assertion NDArray > 2**32` issue that MacOS users have had when generating images larger than 768x768 pixels. Please report back.
+
+As part of the upgrade to `diffusers`, the location of the diffusers-based models has changed from `models/diffusers` to `models/hub`. When you launch InvokeAI for the first time, it will prompt you to OK a one-time move. This should be quick and harmless, but if you have modified your `models/diffusers` directory in some way, for example using symlinks, you may wish to cancel the migration and make appropriate adjustments.
+
+#### New "Invokeai-batch" script
+
+2.3.2 introduces a new command-line only script called
+`invokeai-batch` that can be used to generate hundreds of images from
+prompts and settings that vary systematically. This can be used to try
+the same prompt across multiple combinations of models, steps, CFG
+settings and so forth. It also allows you to template prompts and
+generate a combinatorial list like: ``` a shack in the mountains,
+photograph a shack in the mountains, watercolor a shack in the
+mountains, oil painting a chalet in the mountains, photograph a chalet
+in the mountains, watercolor a chalet in the mountains, oil painting a
+shack in the desert, photograph ...  ```
+
+If you have a system with multiple GPUs, or a single GPU with lots of
+VRAM, you can parallelize generation across the combinatorial set,
+reducing wait times and using your system's resources efficiently
+(make sure you have good GPU cooling).
+
+To try `invokeai-batch` out. Launch the "developer's console" using
+the `invoke` launcher script, or activate the invokeai virtual
+environment manually. From the console, give the command
+`invokeai-batch --help` in order to learn how the script works and
+create your first template file for dynamic prompt generation.
+
+### v2.3.1 <small>(26 February 2023)</small>
+
+This is primarily a bugfix release, but it does provide several new features that will improve the user experience. 
+
+#### Enhanced support for model management
+
+InvokeAI now makes it convenient to add, remove and modify models. You can individually import models that are stored on your local system, scan an entire folder and its subfolders for models and import them automatically, and even directly import models from the internet by providing their download URLs. You also have the option of designating a local folder to scan for new models each time InvokeAI is restarted.
+
+There are three ways of accessing the model management features:
+
+1. ***From the WebUI***, click on the cube to the right of the model selection menu. This will bring up a form that allows you to import models individually from your local disk or scan a directory for models to import.
+
+![image](https://user-images.githubusercontent.com/111189/220638091-918492cc-0719-4194-b033-3741e8289b30.png)
+
+2. **Using the Model Installer App**
+
+Choose option (5) _download and install models_ from the `invoke` launcher script to start a new console-based application for model management. You can use this to select from a curated set of starter models, or import checkpoint, safetensors, and diffusers models from a local disk or the internet. The example below shows importing two checkpoint URLs from popular SD sites and a HuggingFace diffusers model using its Repository ID. It also shows how to designate a folder to be scanned at startup time for new models to import.
+
+Command-line users can start this app using the command `invokeai-model-install`.
+
+![image](https://user-images.githubusercontent.com/111189/220660363-22ff3a2e-8082-410e-a818-d2b3a0529bac.png)
+
+3. **Using the Command Line Client (CLI)**
+
+The `!install_model` and `!convert_model` commands have been enhanced to allow entering of URLs and local directories to scan and import. The first command installs .ckpt and .safetensors files as-is. The second one converts them into the faster diffusers format before installation.
+
+Internally InvokeAI is able to probe the contents of a .ckpt or .safetensors file to distinguish among v1.x, v2.x and inpainting models. This means that you do **not** need to include "inpaint" in your model names to use an inpainting model. Note that Stable Diffusion v2.x models will be autoconverted into a diffusers model the first time you use it.
+
+Please see [INSTALLING MODELS](https://invoke-ai.github.io/InvokeAI/installation/050_INSTALLING_MODELS/) for more information on model management.
+
+#### An Improved Installer Experience
+
+The installer now launches a console-based UI for setting and changing commonly-used startup options:
+
+![image](https://user-images.githubusercontent.com/111189/220644777-3d3a90ca-f9e2-4e6d-93da-cbdd66bf12f3.png)
+
+After selecting the desired options, the installer installs several support models needed by InvokeAI's face reconstruction and upscaling features and then launches the interface for selecting and installing models shown earlier. At any time,  you can edit the startup options by launching `invoke.sh`/`invoke.bat` and entering option (6) _change InvokeAI startup options_
+
+Command-line users can launch the new configure app using `invokeai-configure`.
+
+This release also comes with a renewed updater. To do an update without going through a whole reinstallation, launch `invoke.sh` or `invoke.bat` and choose option (9) _update InvokeAI_ . This will bring you to a screen that prompts you to update to the latest released version, to the most current development version, or any released or unreleased version you choose by selecting the tag or branch of the desired version.
+
+![image](https://user-images.githubusercontent.com/111189/220650124-30a77137-d9cd-406e-a87d-d8283f99a4b3.png)
+
+Command-line users can run this interface by typing `invokeai-configure`
+
+#### Image Symmetry Options
+
+There are now features to generate horizontal and vertical symmetry during generation. The way these work is to wait until a selected step in the generation process and then to turn on a mirror image effect. In addition to generating some cool images, you can also use this to make side-by-side comparisons of how an image will look with more or fewer steps. Access this option from the WebUI by selecting _Symmetry_ from the image generation settings, or within the CLI by using the options `--h_symmetry_time_pct` and `--v_symmetry_time_pct` (these can be abbreviated to `--h_sym` and `--v_sym` like all other options).
+
+![image](https://user-images.githubusercontent.com/111189/220658687-47fd0f2c-7069-4d95-aec9-7196fceb360d.png)
+
+#### A New Unified Canvas Look
+
+This release introduces a beta version of the WebUI Unified Canvas. To try it out, open up the settings dialogue in the WebUI (gear icon) and select _Use Canvas Beta Layout_:
+
+![image](https://user-images.githubusercontent.com/111189/220646958-b7eca95e-dc39-4cd2-b277-63eac98ed446.png)
+
+Refresh the screen and go to to Unified Canvas (left side of screen, third icon from the top). The new layout is designed to provide more space to work in and to keep the image controls close to the image itself:
+
+![image](https://user-images.githubusercontent.com/111189/220647560-4a9265a1-6926-44f9-9d08-e1ef2ce61ff8.png)
+
+#### Model conversion and merging within the WebUI
+
+The WebUI now has an intuitive interface for model merging, as well as for permanent conversion of models from legacy .ckpt/.safetensors formats into diffusers format. These options are also available directly from the `invoke.sh`/`invoke.bat` scripts.
+
+#### An easier way to contribute translations to the WebUI
+
+We have migrated our translation efforts to [Weblate](https://hosted.weblate.org/engage/invokeai/), a FOSS translation product. Maintaining the growing project's translations is now far simpler for the maintainers and community. Please review our brief [translation guide](https://github.com/invoke-ai/InvokeAI/blob/v2.3.1/docs/other/TRANSLATION.md) for more information on how to contribute.
+
+#### Numerous internal bugfixes and performance issues
+
+This releases quashes multiple bugs that were reported in 2.3.0. Major internal changes include upgrading to `diffusers 0.13.0`, and using the `compel` library for prompt parsing. See [Detailed Change Log](#full-change-log) for a detailed list of bugs caught and squished.
+
+#### Summary of InvokeAI command line scripts (all accessible via the launcher menu)
+
+| Command                  | Description                                                         |
+|--------------------------|---------------------------------------------------------------------|
+| `invokeai`               | Command line interface                                              |
+| `invokeai --web`         | Web interface                                                       |
+| `invokeai-model-install` | Model installer with console forms-based front end                  |
+| `invokeai-ti --gui`      | Textual inversion, with a console forms-based front end             |
+| `invokeai-merge --gui`   | Model merging, with a console forms-based front end                 |
+| `invokeai-configure`     | Startup configuration; can also be used to reinstall support models |
+| `invokeai-update`        | InvokeAI software updater                                           |
+
+
 ### v2.3.0 <small>(9 February 2023)</small>
 
 #### Migration to Stable Diffusion `diffusers` models

docs/installation/030_INSTALL_CUDA_AND_ROCM.md

@@ -77,7 +77,7 @@ machine. To test, open up a terminal window and issue the following
 command:
 
 ```
-rocm-smi
+rocminfo
 ```
 
 If you get a table labeled "ROCm System Management Interface" the
@@ -95,9 +95,17 @@ recent version of Ubuntu, 22.04. However, this [community-contributed
 recipe](https://novaspirit.github.io/amdgpu-rocm-ubu22/) is reported
 to work well.
 
-After installation, please run `rocm-smi` a second time to confirm
+After installation, please run `rocminfo` a second time to confirm
 that the driver is present and the GPU is recognized. You may need to
-do a reboot in order to load the driver.
+do a reboot in order to load the driver. In addition, if you see
+errors relating to your username not being a member of the `render`
+group, you may fix this by adding yourself to this group with the command:
+
+```
+sudo usermod -a -G render myUserName
+```
+
+(Thanks to @EgoringKosmos for the usermod recipe.)
 
 ### Linux Install with a ROCm-docker Container
 

docs/installation/050_INSTALLING_MODELS.md

@@ -11,7 +11,7 @@ The model checkpoint files ('\*.ckpt') are the Stable Diffusion
 captioned images gathered from multiple sources.
 
 Originally there was only a single Stable Diffusion weights file,
-which many people named `model.ckpt`. Now there are dozens or more
+which many people named `model.ckpt`. Now there are hundreds
 that have been fine tuned to provide particulary styles, genres, or
 other features. In addition, there are several new formats that
 improve on the original checkpoint format: a `.safetensors` format
@@ -29,9 +29,10 @@ and performance are being made at a rapid pace. Among other features
 is the ability to download and install a `diffusers` model just by
 providing its HuggingFace repository ID.
 
-While InvokeAI will continue to support `.ckpt` and `.safetensors`
+While InvokeAI will continue to support legacy `.ckpt` and `.safetensors`
 models for the near future, these are deprecated and support will
-likely be withdrawn at some point in the not-too-distant future.
+be withdrawn in version 3.0, after which all legacy models will be
+converted into diffusers at the time they are loaded.
 
 This manual will guide you through installing and configuring model
 weight files and converting legacy `.ckpt` and `.safetensors` files
@@ -89,15 +90,18 @@ aware that CIVITAI hosts many models that generate NSFW content.
 !!! note
 
     InvokeAI 2.3.x does not support directly importing and
-    running Stable Diffusion version 2 checkpoint models. You may instead
-    convert them into `diffusers` models using the conversion methods
-    described below.
+    running Stable Diffusion version 2 checkpoint models. If you
+    try to import them, they will be automatically
+    converted into `diffusers` models on the fly. This adds about 20s
+    to loading time. To avoid this overhead, you are encouraged to
+    use one of the conversion methods described below to convert them
+    permanently.
 
 ## Installation
 
 There are multiple ways to install and manage models:
 
-1. The `invokeai-configure` script which will download and install them for you.
+1. The `invokeai-model-install` script which will download and install them for you.
 
 2. The command-line tool (CLI) has commands that allows you to import, configure and modify
    models files.
@@ -105,14 +109,41 @@ There are multiple ways to install and manage models:
 3. The web interface (WebUI) has a GUI for importing and managing
    models.
 
-### Installation via `invokeai-configure`
+### Installation via `invokeai-model-install`
 
-From the `invoke` launcher, choose option (6) "re-run the configure
-script to download new models." This will launch the same script that
-prompted you to select models at install time. You can use this to add
-models that you skipped the first time around. It is all right to
-specify a model that was previously downloaded; the script will just
-confirm that the files are complete.
+From the `invoke` launcher, choose option (5) "Download and install
+models."  This will launch the same script that prompted you to select
+models at install time. You can use this to add models that you
+skipped the first time around. It is all right to specify a model that
+was previously downloaded; the script will just confirm that the files
+are complete.
+
+This script allows you to load 3d party models. Look for a large text
+entry box labeled "IMPORT LOCAL AND REMOTE MODELS." In this box, you
+can cut and paste one or more of any of the following:
+
+1. A URL that points to a downloadable .ckpt or .safetensors file.
+2. A file path pointing to a .ckpt or .safetensors  file.
+3. A diffusers model repo_id (from HuggingFace) in the format
+   "owner/repo_name".
+4. A directory path pointing to a diffusers model directory.
+5. A directory path pointing to a directory containing a bunch of
+   .ckpt and .safetensors files. All will be imported.
+
+You can enter multiple items into the textbox, each one on a separate
+line. You can paste into the textbox using ctrl-shift-V or by dragging
+and dropping a file/directory from the desktop into the box.
+
+The script also lets you designate a directory that will be scanned
+for new model files each time InvokeAI starts up. These models will be
+added automatically.
+
+Lastly, the script gives you a checkbox option to convert legacy models
+into diffusers, or to run the legacy model directly. If you choose to
+convert, the original .ckpt/.safetensors file will **not** be deleted,
+but a new diffusers directory will be created, using twice your disk
+space. However, the diffusers version will load faster, and will be
+compatible with InvokeAI 3.0.
 
 ### Installation via the CLI
 
@@ -144,19 +175,15 @@ invoke> !import_model https://example.org/sd_models/martians.safetensors
 For this to work, the URL must not be password-protected. Otherwise
 you will receive a 404 error.
 
-When you import a legacy model, the CLI will first ask you what type
-of model this is. You can indicate whether it is a model based on
-Stable Diffusion 1.x (1.4 or 1.5), one based on Stable Diffusion 2.x,
-or a 1.x inpainting model. Be careful to indicate the correct model
-type, or it will not load correctly. You can correct the model type
-after the fact using the `!edit_model` command.
-
-The system will then ask you a few other questions about the model,
-including what size image it was trained on (usually 512x512), what
-name and description you wish to use for it, and whether you would
-like to install a custom VAE (variable autoencoder) file for the
-model. For recent models, the answer to the VAE question is usually
-"no," but it won't hurt to answer "yes".
+When you import a legacy model, the CLI will try to figure out what
+type of model it is and select the correct load configuration file.
+However, one thing it can't do is to distinguish between Stable
+Diffusion 2.x models trained on 512x512 vs 768x768 images. In this
+case, the CLI will pop up a menu of choices, asking you to select
+which type of model it is. Please consult the model documentation to
+identify the correct answer, as loading with the wrong configuration
+will lead to black images. You can correct the model type after the
+fact using the `!edit_model` command.
 
 After importing, the model will load. If this is successful, you will
 be asked if you want to keep the model loaded in memory to start
@@ -211,129 +238,6 @@ description for the model, whether to make this the default model that
 is loaded at InvokeAI startup time, and whether to replace its
 VAE. Generally the answer to the latter question is "no".
 
-### Specifying a configuration file for legacy checkpoints
-
-Some checkpoint files come with instructions to use a specific .yaml
-configuration file. For InvokeAI load this file correctly, please put
-the config file in the same directory as the corresponding `.ckpt` or
-`.safetensors` file and make sure the file has the same basename as
-the weights file. Here is an example:
-
-```bash
-wonderful-model-v2.ckpt
-wonderful-model-v2.yaml
-```
-
-Similarly, to use a custom VAE, name the VAE like this:
-
-```bash
-wonderful-model-v2.vae.pt
-```
-
-
-### Converting legacy models into `diffusers`
-
-The CLI `!convert_model` will convert a `.safetensors` or `.ckpt`
-models file into `diffusers` and install it.This will enable the model
-to load and run faster without loss of image quality.
-
-The usage is identical to `!import_model`. You may point the command
-to either a downloaded model file on disk, or to a (non-password
-protected) URL:
-
-```bash
-invoke> !convert_model C:/Users/fred/Downloads/martians.safetensors
-```
-
-After a successful conversion, the CLI will offer you the option of
-deleting the original `.ckpt` or `.safetensors` file.
-
-### Optimizing a previously-installed model
-
-Lastly, if you have previously installed a `.ckpt` or `.safetensors`
-file and wish to convert it into a `diffusers` model, you can do this
-without re-downloading and converting the original file using the
-`!optimize_model` command. Simply pass the short name of an existing
-installed model:
-
-```bash
-invoke> !optimize_model martians-v1.0
-```
-
-The model will be converted into `diffusers` format and replace the
-previously installed version. You will again be offered the
-opportunity to delete the original `.ckpt` or `.safetensors` file.
-
-### Related CLI Commands
-
-There are a whole series of additional model management commands in
-the CLI that you can read about in [Command-Line
-Interface](../features/CLI.md). These include:
-
-* `!models` - List all installed models
-* `!switch <model name>` - Switch to the indicated model
-* `!edit_model <model name>` - Edit the indicated model to change its name, description or other properties
-* `!del_model <model name>` - Delete the indicated model
-
-### Manually editing `configs/models.yaml`
-
-
-If you are comfortable with a text editor then you may simply edit `models.yaml`
-directly.
-
-You will need to download the desired `.ckpt/.safetensors` file and
-place it somewhere on your machine's filesystem. Alternatively, for a
-`diffusers` model, record the repo_id or download the whole model
-directory. Then using a **text** editor (e.g. the Windows Notepad
-application), open the file `configs/models.yaml`, and add a new
-stanza that follows this model:
-
-#### A legacy model
-
-A legacy `.ckpt` or `.safetensors` entry will look like this:
-
-```yaml
-arabian-nights-1.0:
-  description: A great fine-tune in Arabian Nights style
-  weights: ./path/to/arabian-nights-1.0.ckpt
-  config: ./configs/stable-diffusion/v1-inference.yaml
-  format: ckpt
-  width: 512
-  height: 512
-  default: false
-```
-
-Note that `format` is `ckpt` for both `.ckpt` and `.safetensors` files.
-
-#### A diffusers model
-
-A stanza for a `diffusers` model will look like this for a HuggingFace
-model with a repository ID:
-
-```yaml
-arabian-nights-1.1:
-  description: An even better fine-tune of the Arabian Nights
-  repo_id: captahab/arabian-nights-1.1
-  format: diffusers
-  default: true
-```
-
-And for a downloaded directory:
-
-```yaml
-arabian-nights-1.1:
-  description: An even better fine-tune of the Arabian Nights
-  path: /path/to/captahab-arabian-nights-1.1
-  format: diffusers
-  default: true
-```
-
-There is additional syntax for indicating an external VAE to use with
-this model. See `INITIAL_MODELS.yaml` and `models.yaml` for examples.
-
-After you save the modified `models.yaml` file relaunch
-`invokeai`. The new model will now be available for your use.
-
 ### Installation via the WebUI
 
 To access the WebUI Model Manager, click on the button that looks like
@@ -413,3 +317,143 @@ And here is what the same argument looks like in `invokeai.init`:
 --no-nsfw_checker
 --autoconvert /home/fred/stable-diffusion-checkpoints
 ```
+
+### Specifying a configuration file for legacy checkpoints
+
+Some checkpoint files come with instructions to use a specific .yaml
+configuration file. For InvokeAI load this file correctly, please put
+the config file in the same directory as the corresponding `.ckpt` or
+`.safetensors` file and make sure the file has the same basename as
+the model file. Here is an example:
+
+```bash
+wonderful-model-v2.ckpt
+wonderful-model-v2.yaml
+```
+
+This is not needed for `diffusers` models, which come with their own
+pre-packaged configuration.
+
+### Specifying a custom VAE file for legacy checkpoints
+
+To associate a custom VAE with a legacy file, place the VAE file in
+the same directory as the corresponding `.ckpt` or
+`.safetensors` file and make sure the file has the same basename as
+the model file. Use the suffix `.vae.pt` for VAE checkpoint files, and
+`.vae.safetensors` for VAE safetensors files. There is no requirement
+that both the model and the VAE follow the same format.
+
+Example:
+
+```bash
+wonderful-model-v2.pt
+wonderful-model-v2.vae.safetensors
+```
+
+### Converting legacy models into `diffusers`
+
+The CLI `!convert_model` will convert a `.safetensors` or `.ckpt`
+models file into `diffusers` and install it.This will enable the model
+to load and run faster without loss of image quality.
+
+The usage is identical to `!import_model`. You may point the command
+to either a downloaded model file on disk, or to a (non-password
+protected) URL:
+
+```bash
+invoke> !convert_model C:/Users/fred/Downloads/martians.safetensors
+```
+
+After a successful conversion, the CLI will offer you the option of
+deleting the original `.ckpt` or `.safetensors` file.
+
+### Optimizing a previously-installed model
+
+Lastly, if you have previously installed a `.ckpt` or `.safetensors`
+file and wish to convert it into a `diffusers` model, you can do this
+without re-downloading and converting the original file using the
+`!optimize_model` command. Simply pass the short name of an existing
+installed model:
+
+```bash
+invoke> !optimize_model martians-v1.0
+```
+
+The model will be converted into `diffusers` format and replace the
+previously installed version. You will again be offered the
+opportunity to delete the original `.ckpt` or `.safetensors` file.
+
+Alternatively you can use the WebUI's model manager to handle diffusers
+optimization. Select the legacy model you wish to convert, and then
+look for a button labeled "Convert to Diffusers" in the upper right of
+the window.
+
+### Related CLI Commands
+
+There are a whole series of additional model management commands in
+the CLI that you can read about in [Command-Line
+Interface](../features/CLI.md). These include:
+
+* `!models` - List all installed models
+* `!switch <model name>` - Switch to the indicated model
+* `!edit_model <model name>` - Edit the indicated model to change its name, description or other properties
+* `!del_model <model name>` - Delete the indicated model
+
+### Manually editing `configs/models.yaml`
+
+If you are comfortable with a text editor then you may simply edit `models.yaml`
+directly.
+
+You will need to download the desired `.ckpt/.safetensors` file and
+place it somewhere on your machine's filesystem. Alternatively, for a
+`diffusers` model, record the repo_id or download the whole model
+directory. Then using a **text** editor (e.g. the Windows Notepad
+application), open the file `configs/models.yaml`, and add a new
+stanza that follows this model:
+
+#### A legacy model
+
+A legacy `.ckpt` or `.safetensors` entry will look like this:
+
+```yaml
+arabian-nights-1.0:
+  description: A great fine-tune in Arabian Nights style
+  weights: ./path/to/arabian-nights-1.0.ckpt
+  config: ./configs/stable-diffusion/v1-inference.yaml
+  format: ckpt
+  width: 512
+  height: 512
+  default: false
+```
+
+Note that `format` is `ckpt` for both `.ckpt` and `.safetensors` files.
+
+#### A diffusers model
+
+A stanza for a `diffusers` model will look like this for a HuggingFace
+model with a repository ID:
+
+```yaml
+arabian-nights-1.1:
+  description: An even better fine-tune of the Arabian Nights
+  repo_id: captahab/arabian-nights-1.1
+  format: diffusers
+  default: true
+```
+
+And for a downloaded directory:
+
+```yaml
+arabian-nights-1.1:
+  description: An even better fine-tune of the Arabian Nights
+  path: /path/to/captahab-arabian-nights-1.1
+  format: diffusers
+  default: true
+```
+
+There is additional syntax for indicating an external VAE to use with
+this model. See `INITIAL_MODELS.yaml` and `models.yaml` for examples.
+
+After you save the modified `models.yaml` file relaunch
+`invokeai`. The new model will now be available for your use.
+

docs/other/CONTRIBUTORS.md

@@ -23,14 +23,16 @@ We thank them for all of their time and hard work.
 * @damian0815 - Attention Systems and Gameplay Engineer
 * @mauwii (Matthias Wild) - Continuous integration and product maintenance engineer
 * @Netsvetaev (Artur Netsvetaev) - UI/UX Developer
-* @tildebyte - General gadfly and resident (self-appointed) know-it-all
 * @keturn - Lead for Diffusers port
 * @ebr (Eugene Brodsky) - Cloud/DevOps/Sofware engineer; your friendly neighbourhood cluster-autoscaler
 * @jpphoto (Jonathan Pollack) - Inference and rendering engine optimization
 * @genomancer (Gregg Helt) - Model training and merging
+* @gogurtenjoyer - User support and testing
+* @whosawwhatsis - User support and testing
 
 ## **Contributions by**
 
+- [tildebyte](https://github.com/tildebyte)
 - [Sean McLellan](https://github.com/Oceanswave)
 - [Kevin Gibbons](https://github.com/bakkot)
 - [Tesseract Cat](https://github.com/TesseractCat)
@@ -78,6 +80,7 @@ We thank them for all of their time and hard work.
 - [psychedelicious](https://github.com/psychedelicious)
 - [damian0815](https://github.com/damian0815)
 - [Eugene Brodsky](https://github.com/ebr)
+- [Statcomm](https://github.com/statcomm)
 
 ## **Original CompVis Authors**
 

installer/lib/installer.py

@@ -144,8 +144,8 @@ class Installer:
 
             from plumbum import FG, local
 
-            pip = local[get_pip_from_venv(venv_dir)]
-            pip[ "install", "--upgrade", "pip"] & FG
+            python = local[get_python_from_venv(venv_dir)]
+            python[ "-m", "pip", "install", "--upgrade", "pip"] & FG
 
         return venv_dir
 
@@ -242,8 +242,8 @@ class InvokeAiInstance:
         from plumbum import FG, local
 
         # Note that we're installing pinned versions of torch and
-        # torchvision here, which may not correspond to what is
-        # in pyproject.toml. This is a hack to prevent torch 2.0 from
+        # torchvision here, which *should* correspond to what is
+        # in pyproject.toml. This is to prevent torch 2.0 from
         # being installed and immediately uninstalled and replaced with 1.13
         pip = local[self.pip]
 
@@ -252,7 +252,7 @@ class InvokeAiInstance:
                 "install",
                 "--require-virtualenv",
                 "torch~=1.13.1",
-                "torchvision>=0.14.1",
+                "torchvision~=0.14.1",
                 "--force-reinstall",
                 "--find-links" if find_links is not None else None,
                 find_links,
@@ -384,7 +384,7 @@ class InvokeAiInstance:
             os.chmod(dest, 0o0755)
 
         if OS == "Linux":
-            shutil.copy(Path(__file__).parent / '..' / "templates" / "dialogrc", self.runtime / '.dialogrc')
+            shutil.copy(Path(__file__).parents[1] / "templates" / "dialogrc", self.runtime / '.dialogrc')
 
     def update(self):
         pass
@@ -412,6 +412,22 @@ def get_pip_from_venv(venv_path: Path) -> str:
     return str(venv_path.expanduser().resolve() / pip)
 
 
+def get_python_from_venv(venv_path: Path) -> str:
+    """
+    Given a path to a virtual environment, get the absolute path to the `python` executable
+    in a cross-platform fashion. Does not validate that the python executable
+    actually exists in the virtualenv.
+
+    :param venv_path: Path to the virtual environment
+    :type venv_path: Path
+    :return: Absolute path to the python executable
+    :rtype: str
+    """
+
+    python = "Scripts\python.exe" if OS == "Windows" else "bin/python"
+    return str(venv_path.expanduser().resolve() / python)
+
+
 def set_sys_path(venv_path: Path) -> None:
     """
     Given a path to a virtual environment, set the sys.path, in a cross-platform fashion,

invokeai/backend/invoke_ai_web_server.py

@@ -25,15 +25,23 @@ from invokeai.backend.modules.parameters import parameters_to_command
 import invokeai.frontend.dist as frontend
 from ldm.generate import Generate
 from ldm.invoke.args import Args, APP_ID, APP_VERSION, calculate_init_img_hash
-from ldm.invoke.conditioning import get_tokens_for_prompt_object, get_prompt_structure, split_weighted_subprompts, \
-    get_tokenizer
+from ldm.invoke.concepts_lib import HuggingFaceConceptsLibrary
+from ldm.invoke.conditioning import (
+    get_tokens_for_prompt_object,
+    get_prompt_structure,
+    split_weighted_subprompts,
+)
 from ldm.invoke.generator.diffusers_pipeline import PipelineIntermediateState
 from ldm.invoke.generator.inpaint import infill_methods
-from ldm.invoke.globals import Globals, global_converted_ckpts_dir
+from ldm.invoke.globals import (
+    Globals,
+    global_converted_ckpts_dir,
+    global_models_dir,
+)
 from ldm.invoke.pngwriter import PngWriter, retrieve_metadata
 from compel.prompt_parser import Blend
-from ldm.invoke.globals import global_models_dir
 from ldm.invoke.merge_diffusers import merge_diffusion_models
+from ldm.modules.lora_manager import LoraManager
 
 # Loading Arguments
 opt = Args()
@@ -193,8 +201,7 @@ class InvokeAIWebServer:
                 (width, height) = pil_image.size
 
                 thumbnail_path = save_thumbnail(
-                    pil_image, os.path.basename(
-                        file_path), self.thumbnail_image_path
+                    pil_image, os.path.basename(file_path), self.thumbnail_image_path
                 )
 
                 response = {
@@ -224,7 +231,7 @@ class InvokeAIWebServer:
                     server="flask_socketio",
                     width=1600,
                     height=1000,
-                    port=self.port
+                    port=self.port,
                 ).run()
             except KeyboardInterrupt:
                 import sys
@@ -265,16 +272,14 @@ class InvokeAIWebServer:
         # location for "finished" images
         self.result_path = args.outdir
         # temporary path for intermediates
-        self.intermediate_path = os.path.join(
-            self.result_path, "intermediates/")
+        self.intermediate_path = os.path.join(self.result_path, "intermediates/")
         # path for user-uploaded init images and masks
         self.init_image_path = os.path.join(self.result_path, "init-images/")
         self.mask_image_path = os.path.join(self.result_path, "mask-images/")
         # path for temp images e.g. gallery generations which are not committed
         self.temp_image_path = os.path.join(self.result_path, "temp-images/")
         # path for thumbnail images
-        self.thumbnail_image_path = os.path.join(
-            self.result_path, "thumbnails/")
+        self.thumbnail_image_path = os.path.join(self.result_path, "thumbnails/")
         # txt log
         self.log_path = os.path.join(self.result_path, "invoke_log.txt")
         # make all output paths
@@ -299,21 +304,22 @@ class InvokeAIWebServer:
             config["infill_methods"] = infill_methods()
             socketio.emit("systemConfig", config)
 
-        @socketio.on('searchForModels')
+        @socketio.on("searchForModels")
         def handle_search_models(search_folder: str):
             try:
                 if not search_folder:
                     socketio.emit(
                         "foundModels",
-                        {'search_folder': None, 'found_models': None},
+                        {"search_folder": None, "found_models": None},
                     )
                 else:
-                    search_folder, found_models = self.generate.model_manager.search_models(
-                        search_folder)
+                    (
+                        search_folder,
+                        found_models,
+                    ) = self.generate.model_manager.search_models(search_folder)
                     socketio.emit(
                         "foundModels",
-                        {'search_folder': search_folder,
-                            'found_models': found_models},
+                        {"search_folder": search_folder, "found_models": found_models},
                     )
             except Exception as e:
                 self.handle_exceptions(e)
@@ -322,11 +328,11 @@ class InvokeAIWebServer:
         @socketio.on("addNewModel")
         def handle_add_model(new_model_config: dict):
             try:
-                model_name = new_model_config['name']
-                del new_model_config['name']
+                model_name = new_model_config["name"]
+                del new_model_config["name"]
                 model_attributes = new_model_config
-                if len(model_attributes['vae']) == 0:
-                    del model_attributes['vae']
+                if len(model_attributes["vae"]) == 0:
+                    del model_attributes["vae"]
                 update = False
                 current_model_list = self.generate.model_manager.list_models()
                 if model_name in current_model_list:
@@ -335,14 +341,20 @@ class InvokeAIWebServer:
                 print(f">> Adding New Model: {model_name}")
 
                 self.generate.model_manager.add_model(
-                    model_name=model_name, model_attributes=model_attributes, clobber=True)
+                    model_name=model_name,
+                    model_attributes=model_attributes,
+                    clobber=True,
+                )
                 self.generate.model_manager.commit(opt.conf)
 
                 new_model_list = self.generate.model_manager.list_models()
                 socketio.emit(
                     "newModelAdded",
-                    {"new_model_name": model_name,
-                     "model_list": new_model_list, 'update': update},
+                    {
+                        "new_model_name": model_name,
+                        "model_list": new_model_list,
+                        "update": update,
+                    },
                 )
                 print(f">> New Model Added: {model_name}")
             except Exception as e:
@@ -357,8 +369,10 @@ class InvokeAIWebServer:
                 updated_model_list = self.generate.model_manager.list_models()
                 socketio.emit(
                     "modelDeleted",
-                    {"deleted_model_name": model_name,
-                     "model_list": updated_model_list},
+                    {
+                        "deleted_model_name": model_name,
+                        "model_list": updated_model_list,
+                    },
                 )
                 print(f">> Model Deleted: {model_name}")
             except Exception as e:
@@ -383,41 +397,48 @@ class InvokeAIWebServer:
             except Exception as e:
                 self.handle_exceptions(e)
 
-        @socketio.on('convertToDiffusers')
+        @socketio.on("convertToDiffusers")
         def convert_to_diffusers(model_to_convert: dict):
             try:
-                if (model_info := self.generate.model_manager.model_info(model_name=model_to_convert['model_name'])):
-                    if 'weights' in model_info:
-                        ckpt_path = Path(model_info['weights'])
-                        original_config_file = Path(model_info['config'])
-                        model_name = model_to_convert['model_name']
-                        model_description = model_info['description']
+                if model_info := self.generate.model_manager.model_info(
+                    model_name=model_to_convert["model_name"]
+                ):
+                    if "weights" in model_info:
+                        ckpt_path = Path(model_info["weights"])
+                        original_config_file = Path(model_info["config"])
+                        model_name = model_to_convert["model_name"]
+                        model_description = model_info["description"]
                     else:
                         self.socketio.emit(
-                            "error", {"message": "Model is not a valid checkpoint file"})
+                            "error", {"message": "Model is not a valid checkpoint file"}
+                        )
                 else:
                     self.socketio.emit(
-                        "error", {"message": "Could not retrieve model info."})
+                        "error", {"message": "Could not retrieve model info."}
+                    )
 
                 if not ckpt_path.is_absolute():
                     ckpt_path = Path(Globals.root, ckpt_path)
 
                 if original_config_file and not original_config_file.is_absolute():
-                    original_config_file = Path(
-                        Globals.root, original_config_file)
+                    original_config_file = Path(Globals.root, original_config_file)
 
                 diffusers_path = Path(
-                    ckpt_path.parent.absolute(),
-                    f'{model_name}_diffusers'
+                    ckpt_path.parent.absolute(), f"{model_name}_diffusers"
                 )
 
-                if model_to_convert['save_location'] == 'root':
+                if model_to_convert["save_location"] == "root":
                     diffusers_path = Path(
-                        global_converted_ckpts_dir(), f'{model_name}_diffusers')
+                        global_converted_ckpts_dir(), f"{model_name}_diffusers"
+                    )
 
-                if model_to_convert['save_location'] == 'custom' and model_to_convert['custom_location'] is not None:
+                if (
+                    model_to_convert["save_location"] == "custom"
+                    and model_to_convert["custom_location"] is not None
+                ):
                     diffusers_path = Path(
-                        model_to_convert['custom_location'], f'{model_name}_diffusers')
+                        model_to_convert["custom_location"], f"{model_name}_diffusers"
+                    )
 
                 if diffusers_path.exists():
                     shutil.rmtree(diffusers_path)
@@ -435,54 +456,91 @@ class InvokeAIWebServer:
                 new_model_list = self.generate.model_manager.list_models()
                 socketio.emit(
                     "modelConverted",
-                    {"new_model_name": model_name,
-                     "model_list": new_model_list, 'update': True},
+                    {
+                        "new_model_name": model_name,
+                        "model_list": new_model_list,
+                        "update": True,
+                    },
                 )
                 print(f">> Model Converted: {model_name}")
             except Exception as e:
                 self.handle_exceptions(e)
 
-        @socketio.on('mergeDiffusersModels')
+        @socketio.on("mergeDiffusersModels")
         def merge_diffusers_models(model_merge_info: dict):
             try:
-                models_to_merge = model_merge_info['models_to_merge']
+                models_to_merge = model_merge_info["models_to_merge"]
                 model_ids_or_paths = [
-                    self.generate.model_manager.model_name_or_path(x) for x in models_to_merge]
+                    self.generate.model_manager.model_name_or_path(x)
+                    for x in models_to_merge
+                ]
                 merged_pipe = merge_diffusion_models(
-                    model_ids_or_paths, model_merge_info['alpha'], model_merge_info['interp'], model_merge_info['force'])
+                    model_ids_or_paths,
+                    model_merge_info["alpha"],
+                    model_merge_info["interp"],
+                    model_merge_info["force"],
+                )
 
-                dump_path = global_models_dir() / 'merged_models'
-                if model_merge_info['model_merge_save_path'] is not None:
-                    dump_path = Path(model_merge_info['model_merge_save_path'])
+                dump_path = global_models_dir() / "merged_models"
+                if model_merge_info["model_merge_save_path"] is not None:
+                    dump_path = Path(model_merge_info["model_merge_save_path"])
 
                 os.makedirs(dump_path, exist_ok=True)
-                dump_path = dump_path / model_merge_info['merged_model_name']
+                dump_path = dump_path / model_merge_info["merged_model_name"]
                 merged_pipe.save_pretrained(dump_path, safe_serialization=1)
 
                 merged_model_config = dict(
-                    model_name=model_merge_info['merged_model_name'],
+                    model_name=model_merge_info["merged_model_name"],
                     description=f'Merge of models {", ".join(models_to_merge)}',
-                    commit_to_conf=opt.conf
+                    commit_to_conf=opt.conf,
                 )
 
-                if vae := self.generate.model_manager.config[models_to_merge[0]].get("vae", None):
-                    print(
-                        f">> Using configured VAE assigned to {models_to_merge[0]}")
+                if vae := self.generate.model_manager.config[models_to_merge[0]].get(
+                    "vae", None
+                ):
+                    print(f">> Using configured VAE assigned to {models_to_merge[0]}")
                     merged_model_config.update(vae=vae)
 
                 self.generate.model_manager.import_diffuser_model(
-                    dump_path, **merged_model_config)
+                    dump_path, **merged_model_config
+                )
                 new_model_list = self.generate.model_manager.list_models()
 
                 socketio.emit(
                     "modelsMerged",
-                    {"merged_models": models_to_merge,
-                     "merged_model_name": model_merge_info['merged_model_name'],
-                     "model_list": new_model_list, 'update': True},
+                    {
+                        "merged_models": models_to_merge,
+                        "merged_model_name": model_merge_info["merged_model_name"],
+                        "model_list": new_model_list,
+                        "update": True,
+                    },
                 )
                 print(f">> Models Merged: {models_to_merge}")
-                print(
-                    f">> New Model Added: {model_merge_info['merged_model_name']}")
+                print(f">> New Model Added: {model_merge_info['merged_model_name']}")
+            except Exception as e:
+                self.handle_exceptions(e)
+
+        @socketio.on("getLoraModels")
+        def get_lora_models():
+            try:
+                model = self.generate.model
+                lora_mgr = LoraManager(model)
+                loras = lora_mgr.list_compatible_loras()
+                found_loras = []
+                for lora in sorted(loras, key=str.casefold):
+                    found_loras.append({"name":lora,"location":str(loras[lora])})
+                socketio.emit("foundLoras", found_loras)
+            except Exception as e:
+                self.handle_exceptions(e)
+
+        @socketio.on("getTextualInversionTriggers")
+        def get_ti_triggers():
+            try:
+                local_triggers = self.generate.model.textual_inversion_manager.get_all_trigger_strings()
+                locals = [{'name': x} for x in sorted(local_triggers, key=str.casefold)]
+                concepts = HuggingFaceConceptsLibrary().list_concepts(minimum_likes=5)
+                concepts = [{'name': f'<{x}>'} for x in sorted(concepts, key=str.casefold) if f'<{x}>' not in local_triggers]
+                socketio.emit("foundTextualInversionTriggers", {'local_triggers': locals, 'huggingface_concepts': concepts})
             except Exception as e:
                 self.handle_exceptions(e)
 
@@ -500,7 +558,8 @@ class InvokeAIWebServer:
                         os.remove(thumbnail_path)
                     except Exception as e:
                         socketio.emit(
-                            "error", {"message": f"Unable to delete {f}: {str(e)}"})
+                            "error", {"message": f"Unable to delete {f}: {str(e)}"}
+                        )
                         pass
 
                 socketio.emit("tempFolderEmptied")
@@ -511,8 +570,7 @@ class InvokeAIWebServer:
         def save_temp_image_to_gallery(url):
             try:
                 image_path = self.get_image_path_from_url(url)
-                new_path = os.path.join(
-                    self.result_path, os.path.basename(image_path))
+                new_path = os.path.join(self.result_path, os.path.basename(image_path))
                 shutil.copy2(image_path, new_path)
 
                 if os.path.splitext(new_path)[1] == ".png":
@@ -525,8 +583,7 @@ class InvokeAIWebServer:
                 (width, height) = pil_image.size
 
                 thumbnail_path = save_thumbnail(
-                    pil_image, os.path.basename(
-                        new_path), self.thumbnail_image_path
+                    pil_image, os.path.basename(new_path), self.thumbnail_image_path
                 )
 
                 image_array = [
@@ -585,8 +642,7 @@ class InvokeAIWebServer:
                         (width, height) = pil_image.size
 
                         thumbnail_path = save_thumbnail(
-                            pil_image, os.path.basename(
-                                path), self.thumbnail_image_path
+                            pil_image, os.path.basename(path), self.thumbnail_image_path
                         )
 
                         image_array.append(
@@ -605,7 +661,8 @@ class InvokeAIWebServer:
                         )
                     except Exception as e:
                         socketio.emit(
-                            "error", {"message": f"Unable to load {path}: {str(e)}"})
+                            "error", {"message": f"Unable to load {path}: {str(e)}"}
+                        )
                         pass
 
                 socketio.emit(
@@ -655,8 +712,7 @@ class InvokeAIWebServer:
                         (width, height) = pil_image.size
 
                         thumbnail_path = save_thumbnail(
-                            pil_image, os.path.basename(
-                                path), self.thumbnail_image_path
+                            pil_image, os.path.basename(path), self.thumbnail_image_path
                         )
 
                         image_array.append(
@@ -676,7 +732,8 @@ class InvokeAIWebServer:
                     except Exception as e:
                         print(f">> Unable to load {path}")
                         socketio.emit(
-                            "error", {"message": f"Unable to load {path}: {str(e)}"})
+                            "error", {"message": f"Unable to load {path}: {str(e)}"}
+                        )
                         pass
 
                 socketio.emit(
@@ -710,10 +767,9 @@ class InvokeAIWebServer:
                         printable_parameters["init_mask"][:64] + "..."
                     )
 
-                print(
-                    f'\n>> Image Generation Parameters:\n\n{printable_parameters}\n')
-                print(f'>> ESRGAN Parameters: {esrgan_parameters}')
-                print(f'>> Facetool Parameters: {facetool_parameters}')
+                print(f"\n>> Image Generation Parameters:\n\n{printable_parameters}\n")
+                print(f">> ESRGAN Parameters: {esrgan_parameters}")
+                print(f">> Facetool Parameters: {facetool_parameters}")
 
                 self.generate_images(
                     generation_parameters,
@@ -750,11 +806,9 @@ class InvokeAIWebServer:
                 if postprocessing_parameters["type"] == "esrgan":
                     progress.set_current_status("common.statusUpscalingESRGAN")
                 elif postprocessing_parameters["type"] == "gfpgan":
-                    progress.set_current_status(
-                        "common.statusRestoringFacesGFPGAN")
+                    progress.set_current_status("common.statusRestoringFacesGFPGAN")
                 elif postprocessing_parameters["type"] == "codeformer":
-                    progress.set_current_status(
-                        "common.statusRestoringFacesCodeFormer")
+                    progress.set_current_status("common.statusRestoringFacesCodeFormer")
 
                 socketio.emit("progressUpdate", progress.to_formatted_dict())
                 eventlet.sleep(0)
@@ -919,8 +973,7 @@ class InvokeAIWebServer:
 
                 init_img_url = generation_parameters["init_img"]
 
-                original_bounding_box = generation_parameters["bounding_box"].copy(
-                )
+                original_bounding_box = generation_parameters["bounding_box"].copy()
 
                 initial_image = dataURL_to_image(
                     generation_parameters["init_img"]
@@ -997,8 +1050,9 @@ class InvokeAIWebServer:
             elif generation_parameters["generation_mode"] == "img2img":
                 init_img_url = generation_parameters["init_img"]
                 init_img_path = self.get_image_path_from_url(init_img_url)
-                generation_parameters["init_img"] = Image.open(
-                    init_img_path).convert('RGB')
+                generation_parameters["init_img"] = Image.open(init_img_path).convert(
+                    "RGB"
+                )
 
             def image_progress(sample, step):
                 if self.canceled.is_set():
@@ -1058,8 +1112,7 @@ class InvokeAIWebServer:
                     )
 
                 if generation_parameters["progress_latents"]:
-                    image = self.generate.sample_to_lowres_estimated_image(
-                        sample)
+                    image = self.generate.sample_to_lowres_estimated_image(sample)
                     (width, height) = image.size
                     width *= 8
                     height *= 8
@@ -1078,8 +1131,7 @@ class InvokeAIWebServer:
                         },
                     )
 
-                self.socketio.emit(
-                    "progressUpdate", progress.to_formatted_dict())
+                self.socketio.emit("progressUpdate", progress.to_formatted_dict())
                 eventlet.sleep(0)
 
             def image_done(image, seed, first_seed, attention_maps_image=None):
@@ -1106,8 +1158,7 @@ class InvokeAIWebServer:
 
                 progress.set_current_status("common.statusGenerationComplete")
 
-                self.socketio.emit(
-                    "progressUpdate", progress.to_formatted_dict())
+                self.socketio.emit("progressUpdate", progress.to_formatted_dict())
                 eventlet.sleep(0)
 
                 all_parameters = generation_parameters
@@ -1118,8 +1169,7 @@ class InvokeAIWebServer:
                     and all_parameters["variation_amount"] > 0
                 ):
                     first_seed = first_seed or seed
-                    this_variation = [
-                        [seed, all_parameters["variation_amount"]]]
+                    this_variation = [[seed, all_parameters["variation_amount"]]]
                     all_parameters["with_variations"] = (
                         prior_variations + this_variation
                     )
@@ -1135,14 +1185,13 @@ class InvokeAIWebServer:
                 if esrgan_parameters:
                     progress.set_current_status("common.statusUpscaling")
                     progress.set_current_status_has_steps(False)
-                    self.socketio.emit(
-                        "progressUpdate", progress.to_formatted_dict())
+                    self.socketio.emit("progressUpdate", progress.to_formatted_dict())
                     eventlet.sleep(0)
 
                     image = self.esrgan.process(
                         image=image,
                         upsampler_scale=esrgan_parameters["level"],
-                        denoise_str=esrgan_parameters['denoise_str'],
+                        denoise_str=esrgan_parameters["denoise_str"],
                         strength=esrgan_parameters["strength"],
                         seed=seed,
                     )
@@ -1150,7 +1199,7 @@ class InvokeAIWebServer:
                     postprocessing = True
                     all_parameters["upscale"] = [
                         esrgan_parameters["level"],
-                        esrgan_parameters['denoise_str'],
+                        esrgan_parameters["denoise_str"],
                         esrgan_parameters["strength"],
                     ]
 
@@ -1159,15 +1208,14 @@ class InvokeAIWebServer:
 
                 if facetool_parameters:
                     if facetool_parameters["type"] == "gfpgan":
-                        progress.set_current_status(
-                            "common.statusRestoringFacesGFPGAN")
+                        progress.set_current_status("common.statusRestoringFacesGFPGAN")
                     elif facetool_parameters["type"] == "codeformer":
                         progress.set_current_status(
-                            "common.statusRestoringFacesCodeFormer")
+                            "common.statusRestoringFacesCodeFormer"
+                        )
 
                     progress.set_current_status_has_steps(False)
-                    self.socketio.emit(
-                        "progressUpdate", progress.to_formatted_dict())
+                    self.socketio.emit("progressUpdate", progress.to_formatted_dict())
                     eventlet.sleep(0)
 
                     if facetool_parameters["type"] == "gfpgan":
@@ -1197,8 +1245,7 @@ class InvokeAIWebServer:
                     all_parameters["facetool_type"] = facetool_parameters["type"]
 
                 progress.set_current_status("common.statusSavingImage")
-                self.socketio.emit(
-                    "progressUpdate", progress.to_formatted_dict())
+                self.socketio.emit("progressUpdate", progress.to_formatted_dict())
                 eventlet.sleep(0)
 
                 # restore the stashed URLS and discard the paths, we are about to send the result to client
@@ -1215,8 +1262,7 @@ class InvokeAIWebServer:
                 if generation_parameters["generation_mode"] == "unifiedCanvas":
                     all_parameters["bounding_box"] = original_bounding_box
 
-                metadata = self.parameters_to_generated_image_metadata(
-                    all_parameters)
+                metadata = self.parameters_to_generated_image_metadata(all_parameters)
 
                 command = parameters_to_command(all_parameters)
 
@@ -1246,22 +1292,27 @@ class InvokeAIWebServer:
 
                 if progress.total_iterations > progress.current_iteration:
                     progress.set_current_step(1)
-                    progress.set_current_status(
-                        "common.statusIterationComplete")
+                    progress.set_current_status("common.statusIterationComplete")
                     progress.set_current_status_has_steps(False)
                 else:
                     progress.mark_complete()
 
-                self.socketio.emit(
-                    "progressUpdate", progress.to_formatted_dict())
+                self.socketio.emit("progressUpdate", progress.to_formatted_dict())
                 eventlet.sleep(0)
 
-                parsed_prompt, _ = get_prompt_structure(
-                    generation_parameters["prompt"])
-                tokens = None if type(parsed_prompt) is Blend else \
-                    get_tokens_for_prompt_object(get_tokenizer(self.generate.model), parsed_prompt)
-                attention_maps_image_base64_url = None if attention_maps_image is None \
+                parsed_prompt, _ = get_prompt_structure(generation_parameters["prompt"])
+                tokens = (
+                    None
+                    if type(parsed_prompt) is Blend
+                    else get_tokens_for_prompt_object(
+                        self.generate.model.tokenizer, parsed_prompt
+                    )
+                )
+                attention_maps_image_base64_url = (
+                    None
+                    if attention_maps_image is None
                     else image_to_dataURL(attention_maps_image)
+                )
 
                 self.socketio.emit(
                     "generationResult",
@@ -1293,7 +1344,7 @@ class InvokeAIWebServer:
             self.generate.prompt2image(
                 **generation_parameters,
                 step_callback=diffusers_step_callback_adapter,
-                image_callback=image_done
+                image_callback=image_done,
             )
 
         except KeyboardInterrupt:
@@ -1416,8 +1467,7 @@ class InvokeAIWebServer:
         self, parameters, original_image_path
     ):
         try:
-            current_metadata = retrieve_metadata(
-                original_image_path)["sd-metadata"]
+            current_metadata = retrieve_metadata(original_image_path)["sd-metadata"]
             postprocessing_metadata = {}
 
             """
@@ -1457,8 +1507,7 @@ class InvokeAIWebServer:
                     postprocessing_metadata
                 )
             else:
-                current_metadata["image"]["postprocessing"] = [
-                    postprocessing_metadata]
+                current_metadata["image"]["postprocessing"] = [postprocessing_metadata]
 
             return current_metadata
 
@@ -1554,8 +1603,7 @@ class InvokeAIWebServer:
                 )
             elif "thumbnails" in url:
                 return os.path.abspath(
-                    os.path.join(self.thumbnail_image_path,
-                                 os.path.basename(url))
+                    os.path.join(self.thumbnail_image_path, os.path.basename(url))
                 )
             else:
                 return os.path.abspath(
@@ -1601,7 +1649,7 @@ class InvokeAIWebServer:
         except Exception as e:
             self.handle_exceptions(e)
 
-    def handle_exceptions(self, exception, emit_key: str = 'error'):
+    def handle_exceptions(self, exception, emit_key: str = "error"):
         self.socketio.emit(emit_key, {"message": (str(exception))})
         print("\n")
         traceback.print_exc()

invokeai/configs/INITIAL_MODELS.yaml

@@ -80,7 +80,8 @@ trinart-2.0:
      repo_id: stabilityai/sd-vae-ft-mse
    recommended: False
 waifu-diffusion-1.4:
-   description: An SD-1.5 model trained on 680k anime/manga-style images (2.13 GB)
+   description: An SD-2.1 model trained on 5.4M anime/manga-style images (4.27 GB)
+   revision: main
    repo_id: hakurei/waifu-diffusion
    format: diffusers
    vae:

invokeai/frontend/dist/index.html

@@ -5,8 +5,8 @@
     <meta name="viewport" content="width=device-width, initial-scale=1.0" />
     <title>InvokeAI - A Stable Diffusion Toolkit</title>
     <link rel="shortcut icon" type="icon" href="./assets/favicon-0d253ced.ico" />
-    <script type="module" crossorigin src="./assets/index-c09cf9ca.js"></script>
-    <link rel="stylesheet" href="./assets/index-14cb2922.css">
+    <script type="module" crossorigin src="./assets/index-b12e648e.js"></script>
+    <link rel="stylesheet" href="./assets/index-2ab0eb58.css">
   </head>
 
   <body>

invokeai/frontend/dist/locales/en.json

@@ -327,6 +327,13 @@
         "addModel": "Add Model",
         "updateModel": "Update Model",
         "availableModels": "Available Models",
+        "addLora": "Add Lora",
+        "clearLoras": "Clear Loras",
+        "noLoraModels": "No Loras Found",
+        "addTextualInversionTrigger": "Add Textual Inversion",
+        "addTIToNegative": "Add To Negative",
+        "clearTextualInversions": "Clear Textual Inversions",
+        "noTextualInversionTriggers": "No Textual Inversions Found",
         "search": "Search",
         "load": "Load",
         "active": "active",
@@ -483,6 +490,7 @@
         "useCanvasBeta": "Use Canvas Beta Layout",
         "enableImageDebugging": "Enable Image Debugging",
         "useSlidersForAll": "Use Sliders For All Options",
+        "showHuggingFaceConcepts": "Show Textual Inversions from HF Concepts Library",
         "resetWebUI": "Reset Web UI",
         "resetWebUIDesc1": "Resetting the web UI only resets the browser's local cache of your images and remembered settings. It does not delete any images from disk.",
         "resetWebUIDesc2": "If images aren't showing up in the gallery or something else isn't working, please try resetting before submitting an issue on GitHub.",

invokeai/frontend/package.json

@@ -15,8 +15,8 @@
     "postinstall": "patch-package"
   },
   "dependencies": {
-    "@chakra-ui/icons": "^2.0.17",
-    "@chakra-ui/react": "^2.5.1",
+    "@chakra-ui/icons": "^2.0.18",
+    "@chakra-ui/react": "^2.5.5",
     "@emotion/cache": "^11.10.5",
     "@emotion/react": "^11.10.6",
     "@emotion/styled": "^11.10.6",
@@ -52,6 +52,7 @@
     "redux-persist": "^6.0.0",
     "socket.io": "^4.6.0",
     "socket.io-client": "^4.6.0",
+    "typescript": "^5.0.3",
     "use-image": "^1.1.0",
     "uuid": "^9.0.0",
     "yarn": "^1.22.19"
@@ -61,8 +62,8 @@
     "@types/react": "^18.0.28",
     "@types/react-dom": "^18.0.11",
     "@types/react-transition-group": "^4.4.5",
-    "@typescript-eslint/eslint-plugin": "^5.52.0",
-    "@typescript-eslint/parser": "^5.52.0",
+    "@typescript-eslint/eslint-plugin": "^5.57.0",
+    "@typescript-eslint/parser": "^5.57.0",
     "babel-plugin-transform-imports": "^2.0.0",
     "eslint": "^8.34.0",
     "eslint-config-prettier": "^8.6.0",

invokeai/frontend/public/locales/en.json

@@ -327,6 +327,13 @@
         "addModel": "Add Model",
         "updateModel": "Update Model",
         "availableModels": "Available Models",
+        "addLora": "Add Lora",
+        "clearLoras": "Clear Loras",
+        "noLoraModels": "No Loras Found",
+        "addTextualInversionTrigger": "Add Textual Inversion",
+        "addTIToNegative": "Add To Negative",
+        "clearTextualInversions": "Clear Textual Inversions",
+        "noTextualInversionTriggers": "No Textual Inversions Found",
         "search": "Search",
         "load": "Load",
         "active": "active",
@@ -483,6 +490,7 @@
         "useCanvasBeta": "Use Canvas Beta Layout",
         "enableImageDebugging": "Enable Image Debugging",
         "useSlidersForAll": "Use Sliders For All Options",
+        "showHuggingFaceConcepts": "Show Textual Inversions from HF Concepts Library",
         "resetWebUI": "Reset Web UI",
         "resetWebUIDesc1": "Resetting the web UI only resets the browser's local cache of your images and remembered settings. It does not delete any images from disk.",
         "resetWebUIDesc2": "If images aren't showing up in the gallery or something else isn't working, please try resetting before submitting an issue on GitHub.",

invokeai/frontend/src/app/invokeai.d.ts

@@ -271,6 +271,23 @@ export declare type FoundModelResponse = {
   found_models: FoundModel[];
 };
 
+export declare type FoundLora = {
+  name: string;
+  location: string;
+};
+
+export declare type FoundTextualInversionTriggers = {
+  name: string;
+  location: string;
+};
+
+export declare type FoundLorasRsponse = FoundLora[];
+
+export declare type FoundTextualInversionTriggersResponse = {
+  local_triggers: FoundTextualInversionTriggers[];
+  huggingface_concepts: FoundTextualInversionTriggers[];
+};
+
 export declare type SystemStatusResponse = SystemStatus;
 
 export declare type SystemConfigResponse = SystemConfig;

invokeai/frontend/src/app/socketio/actions.ts

@@ -52,6 +52,12 @@ export const requestModelChange = createAction<string>(
   'socketio/requestModelChange'
 );
 
+export const getLoraModels = createAction<undefined>('socketio/getLoraModels');
+
+export const getTextualInversionTriggers = createAction<undefined>(
+  'socketio/getTextualInversionTriggers'
+);
+
 export const saveStagingAreaImageToGallery = createAction<string>(
   'socketio/saveStagingAreaImageToGallery'
 );

invokeai/frontend/src/app/socketio/emitters.ts

@@ -196,6 +196,12 @@ const makeSocketIOEmitters = (
       dispatch(modelChangeRequested());
       socketio.emit('requestModelChange', modelName);
     },
+    emitGetLoraModels: () => {
+      socketio.emit('getLoraModels');
+    },
+    emitGetTextualInversionTriggers: () => {
+      socketio.emit('getTextualInversionTriggers');
+    },
     emitSaveStagingAreaImageToGallery: (url: string) => {
       socketio.emit('requestSaveStagingAreaImageToGallery', url);
     },

invokeai/frontend/src/app/socketio/listeners.ts

@@ -11,6 +11,7 @@ import {
   errorOccurred,
   processingCanceled,
   setCurrentStatus,
+  setFoundLoras,
   setFoundModels,
   setIsCancelable,
   setIsConnected,
@@ -19,6 +20,8 @@ import {
   setSearchFolder,
   setSystemConfig,
   setSystemStatus,
+  setFoundLocalTextualInversionTriggers,
+  setFoundHuggingFaceTextualInversionTriggers,
 } from 'features/system/store/systemSlice';
 
 import {
@@ -30,12 +33,18 @@ import {
   setIntermediateImage,
 } from 'features/gallery/store/gallerySlice';
 
+import {
+  getLoraModels,
+  getTextualInversionTriggers,
+} from 'app/socketio/actions';
 import type { RootState } from 'app/store';
 import { addImageToStagingArea } from 'features/canvas/store/canvasSlice';
 import {
   clearInitialImage,
+  setHuggingFaceTextualInversionConcepts,
   setInfillMethod,
   setInitialImage,
+  setLocalTextualInversionTriggers,
   setMaskPath,
 } from 'features/parameters/store/generationSlice';
 import { tabMap } from 'features/ui/store/tabMap';
@@ -458,6 +467,8 @@ const makeSocketIOListeners = (
       const { model_name, model_list } = data;
       dispatch(setModelList(model_list));
       dispatch(setCurrentStatus(i18n.t('common.statusModelChanged')));
+      dispatch(getLoraModels());
+      dispatch(getTextualInversionTriggers());
       dispatch(setIsProcessing(false));
       dispatch(setIsCancelable(true));
       dispatch(
@@ -482,6 +493,37 @@ const makeSocketIOListeners = (
         })
       );
     },
+    onFoundLoras: (data: InvokeAI.FoundLorasRsponse) => {
+      dispatch(setFoundLoras(data));
+    },
+    onFoundTextualInversionTriggers: (
+      data: InvokeAI.FoundTextualInversionTriggersResponse
+    ) => {
+      const localTriggers = data.local_triggers;
+      const huggingFaceConcepts = data.huggingface_concepts;
+
+      dispatch(setFoundLocalTextualInversionTriggers(localTriggers));
+      dispatch(
+        setFoundHuggingFaceTextualInversionTriggers(huggingFaceConcepts)
+      );
+
+      // Assign Local TI's
+      const foundLocalTINames: string[] = [];
+      localTriggers.forEach((textualInversion) => {
+        foundLocalTINames.push(textualInversion.name);
+      });
+      dispatch(setLocalTextualInversionTriggers(foundLocalTINames));
+
+      // Assign HuggingFace Concepts
+      const foundHuggingFaceConceptNames: string[] = [];
+      huggingFaceConcepts.forEach((concept) => {
+        foundHuggingFaceConceptNames.push(concept.name);
+      });
+      dispatch(
+        setHuggingFaceTextualInversionConcepts(foundHuggingFaceConceptNames)
+      );
+    },
+
     onTempFolderEmptied: () => {
       dispatch(
         addToast({

invokeai/frontend/src/app/socketio/middleware.ts

@@ -51,6 +51,8 @@ export const socketioMiddleware = () => {
       onModelConverted,
       onModelsMerged,
       onModelChangeFailed,
+      onFoundLoras,
+      onFoundTextualInversionTriggers,
       onTempFolderEmptied,
     } = makeSocketIOListeners(store);
 
@@ -69,6 +71,8 @@ export const socketioMiddleware = () => {
       emitConvertToDiffusers,
       emitMergeDiffusersModels,
       emitRequestModelChange,
+      emitGetLoraModels,
+      emitGetTextualInversionTriggers,
       emitSaveStagingAreaImageToGallery,
       emitRequestEmptyTempFolder,
     } = makeSocketIOEmitters(store, socketio);
@@ -145,6 +149,17 @@ export const socketioMiddleware = () => {
         onModelChangeFailed(data);
       });
 
+      socketio.on('foundLoras', (data: InvokeAI.FoundLorasRsponse) => {
+        onFoundLoras(data);
+      });
+
+      socketio.on(
+        'foundTextualInversionTriggers',
+        (data: InvokeAI.FoundTextualInversionTriggersResponse) => {
+          onFoundTextualInversionTriggers(data);
+        }
+      );
+
       socketio.on('tempFolderEmptied', () => {
         onTempFolderEmptied();
       });
@@ -226,6 +241,16 @@ export const socketioMiddleware = () => {
         break;
       }
 
+      case 'socketio/getLoraModels': {
+        emitGetLoraModels();
+        break;
+      }
+
+      case 'socketio/getTextualInversionTriggers': {
+        emitGetTextualInversionTriggers();
+        break;
+      }
+
       case 'socketio/saveStagingAreaImageToGallery': {
         emitSaveStagingAreaImageToGallery(action.payload);
         break;

invokeai/frontend/src/common/components/IAISimpleMenu.tsx

@@ -7,13 +7,14 @@ import {
   MenuButtonProps,
   MenuListProps,
   MenuItemProps,
+  Text,
 } from '@chakra-ui/react';
 import { MouseEventHandler, ReactNode } from 'react';
 import { MdArrowDropDown, MdArrowDropUp } from 'react-icons/md';
 import IAIButton from './IAIButton';
 import IAIIconButton from './IAIIconButton';
 
-interface IAIMenuItem {
+export interface IAIMenuItem {
   item: ReactNode | string;
   onClick: MouseEventHandler<HTMLButtonElement> | undefined;
 }
@@ -43,6 +44,7 @@ export default function IAISimpleMenu(props: IAIMenuProps) {
 
   const renderMenuItems = () => {
     const menuItemsToRender: ReactNode[] = [];
+
     menuItems.forEach((menuItem, index) => {
       menuItemsToRender.push(
         <MenuItem
@@ -82,12 +84,17 @@ export default function IAISimpleMenu(props: IAIMenuProps) {
             fontSize="1.5rem"
             {...menuButtonProps}
           >
-            {menuType === 'regular' && buttonText}
+            {menuType === 'regular' && (
+              <Text fontSize="0.9rem">{buttonText}</Text>
+            )}
           </MenuButton>
           <MenuList
             zIndex={15}
             padding={0}
             borderRadius="0.5rem"
+            overflow="scroll"
+            maxWidth={'22.5rem'}
+            maxHeight={500}
             backgroundColor="var(--background-color-secondary)"
             color="var(--text-color-secondary)"
             borderColor="var(--border-color)"

invokeai/frontend/src/features/parameters/components/MainParameters/MainWidth.tsx

@@ -34,7 +34,6 @@ export default function MainWidth() {
       withSliderMarks
       sliderMarkRightOffset={-8}
       inputWidth="6.2rem"
-      inputReadOnly
       sliderNumberInputProps={{ max: 15360 }}
     />
   ) : (

invokeai/frontend/src/features/parameters/components/PromptInput/Extras/LoraManager/LoraManager.tsx

@@ -0,0 +1,86 @@
+import { Box, Flex } from '@chakra-ui/react';
+import { getLoraModels } from 'app/socketio/actions';
+import { useAppDispatch, useAppSelector } from 'app/storeHooks';
+import IAIIconButton from 'common/components/IAIIconButton';
+import IAISimpleMenu, { IAIMenuItem } from 'common/components/IAISimpleMenu';
+import {
+  setClearLoras,
+  setLorasInUse,
+} from 'features/parameters/store/generationSlice';
+import { useEffect } from 'react';
+import { useTranslation } from 'react-i18next';
+import { MdClear } from 'react-icons/md';
+
+export default function LoraManager() {
+  const dispatch = useAppDispatch();
+  const foundLoras = useAppSelector((state) => state.system.foundLoras);
+  const lorasInUse = useAppSelector((state) => state.generation.lorasInUse);
+
+  const { t } = useTranslation();
+
+  const handleLora = (lora: string) => {
+    dispatch(setLorasInUse(lora));
+  };
+
+  useEffect(() => {
+    dispatch(getLoraModels());
+  }, [dispatch]);
+
+  const renderLoraOption = (lora: string) => {
+    const thisloraExists = lorasInUse.includes(lora);
+    const loraExistsStyle = {
+      fontWeight: 'bold',
+      color: 'var(--context-menu-active-item)',
+    };
+    return <Box style={thisloraExists ? loraExistsStyle : {}}>{lora}</Box>;
+  };
+
+  const numOfActiveLoras = () => {
+    const foundLoraNames: string[] = [];
+    foundLoras?.forEach((lora) => {
+      foundLoraNames.push(lora.name);
+    });
+    return foundLoraNames.filter((lora) => lorasInUse.includes(lora)).length;
+  };
+
+  const makeLoraItems = () => {
+    const lorasFound: IAIMenuItem[] = [];
+    foundLoras?.forEach((lora) => {
+      if (lora.name !== ' ') {
+        const newLoraItem: IAIMenuItem = {
+          item: renderLoraOption(lora.name),
+          onClick: () => handleLora(lora.name),
+        };
+        lorasFound.push(newLoraItem);
+      }
+    });
+    return lorasFound;
+  };
+
+  return foundLoras && foundLoras?.length > 0 ? (
+    <Flex columnGap={2}>
+      <IAISimpleMenu
+        menuItems={makeLoraItems()}
+        menuType="regular"
+        buttonText={`${t('modelManager.addLora')} (${numOfActiveLoras()})`}
+        menuButtonProps={{ width: '100%', padding: '0 1rem' }}
+      />
+      <IAIIconButton
+        icon={<MdClear />}
+        tooltip={t('modelManager.clearLoras')}
+        aria-label={t('modelManager.clearLoras')}
+        onClick={() => dispatch(setClearLoras())}
+      />
+    </Flex>
+  ) : (
+    <Box
+      background="var(--btn-base-color)"
+      padding={2}
+      textAlign="center"
+      borderRadius={4}
+      fontWeight="bold"
+    >
+      {t('modelManager.noLoraModels')}
+    </Box>
+  );
+}

invokeai/frontend/src/features/parameters/components/PromptInput/Extras/PromptExtras.tsx

@@ -0,0 +1,12 @@
+import { Flex } from '@chakra-ui/react';
+import LoraManager from './LoraManager/LoraManager';
+import TextualInversionManager from './TextualInversionManager/TextualInversionManager';
+
+export default function PromptExtras() {
+  return (
+    <Flex flexDir="column" rowGap={2}>
+      <LoraManager />
+      <TextualInversionManager />
+    </Flex>
+  );
+}

invokeai/frontend/src/features/parameters/components/PromptInput/Extras/TextualInversionManager/TextualInversionManager.tsx

@@ -0,0 +1,163 @@
+import { Box, Flex } from '@chakra-ui/react';
+import { getTextualInversionTriggers } from 'app/socketio/actions';
+import { RootState } from 'app/store';
+import { useAppDispatch, useAppSelector } from 'app/storeHooks';
+import IAIIconButton from 'common/components/IAIIconButton';
+import IAISimpleMenu, { IAIMenuItem } from 'common/components/IAISimpleMenu';
+import {
+  setAddTIToNegative,
+  setClearTextualInversions,
+  setTextualInversionsInUse,
+} from 'features/parameters/store/generationSlice';
+import { useEffect } from 'react';
+import { useTranslation } from 'react-i18next';
+import { MdArrowDownward, MdClear } from 'react-icons/md';
+
+export default function TextualInversionManager() {
+  const dispatch = useAppDispatch();
+  const textualInversionsInUse = useAppSelector(
+    (state: RootState) => state.generation.textualInversionsInUse
+  );
+
+  const negativeTextualInversionsInUse = useAppSelector(
+    (state: RootState) => state.generation.negativeTextualInversionsInUse
+  );
+
+  const foundLocalTextualInversionTriggers = useAppSelector(
+    (state) => state.system.foundLocalTextualInversionTriggers
+  );
+  const foundHuggingFaceTextualInversionTriggers = useAppSelector(
+    (state) => state.system.foundHuggingFaceTextualInversionTriggers
+  );
+
+  const localTextualInversionTriggers = useAppSelector(
+    (state) => state.generation.localTextualInversionTriggers
+  );
+
+  const huggingFaceTextualInversionConcepts = useAppSelector(
+    (state) => state.generation.huggingFaceTextualInversionConcepts
+  );
+
+  const shouldShowHuggingFaceConcepts = useAppSelector(
+    (state) => state.ui.shouldShowHuggingFaceConcepts
+  );
+
+  const addTIToNegative = useAppSelector(
+    (state) => state.generation.addTIToNegative
+  );
+
+  const { t } = useTranslation();
+
+  useEffect(() => {
+    dispatch(getTextualInversionTriggers());
+  }, [dispatch]);
+
+  const handleTextualInversion = (textual_inversion: string) => {
+    dispatch(setTextualInversionsInUse(textual_inversion));
+  };
+
+  const TIPip = ({ color }: { color: string }) => {
+    return (
+      <Box width={2} height={2} borderRadius={9999} backgroundColor={color}>
+        {' '}
+      </Box>
+    );
+  };
+
+  const renderTextualInversionOption = (textual_inversion: string) => {
+    return (
+      <Flex alignItems="center" columnGap={1}>
+        {textual_inversion}
+        {textualInversionsInUse.includes(textual_inversion) && (
+          <TIPip color="var(--context-menu-active-item)" />
+        )}
+        {negativeTextualInversionsInUse.includes(textual_inversion) && (
+          <TIPip color="var(--status-bad-color)" />
+        )}
+      </Flex>
+    );
+  };
+
+  const numOfActiveTextualInversions = () => {
+    const allTextualInversions = localTextualInversionTriggers.concat(
+      huggingFaceTextualInversionConcepts
+    );
+    return allTextualInversions.filter(
+      (ti) =>
+        textualInversionsInUse.includes(ti) ||
+        negativeTextualInversionsInUse.includes(ti)
+    ).length;
+  };
+
+  const makeTextualInversionItems = () => {
+    const textualInversionsFound: IAIMenuItem[] = [];
+    foundLocalTextualInversionTriggers?.forEach((textualInversion) => {
+      if (textualInversion.name !== ' ') {
+        const newTextualInversionItem: IAIMenuItem = {
+          item: renderTextualInversionOption(textualInversion.name),
+          onClick: () => handleTextualInversion(textualInversion.name),
+        };
+        textualInversionsFound.push(newTextualInversionItem);
+      }
+    });
+
+    if (shouldShowHuggingFaceConcepts) {
+      foundHuggingFaceTextualInversionTriggers?.forEach((textualInversion) => {
+        if (textualInversion.name !== ' ') {
+          const newTextualInversionItem: IAIMenuItem = {
+            item: renderTextualInversionOption(textualInversion.name),
+            onClick: () => handleTextualInversion(textualInversion.name),
+          };
+          textualInversionsFound.push(newTextualInversionItem);
+        }
+      });
+    }
+
+    return textualInversionsFound;
+  };
+
+  return foundLocalTextualInversionTriggers &&
+    (foundLocalTextualInversionTriggers?.length > 0 ||
+      (foundHuggingFaceTextualInversionTriggers &&
+        foundHuggingFaceTextualInversionTriggers?.length > 0 &&
+        shouldShowHuggingFaceConcepts)) ? (
+    <Flex columnGap={2}>
+      <IAISimpleMenu
+        menuItems={makeTextualInversionItems()}
+        menuType="regular"
+        buttonText={`${t(
+          'modelManager.addTextualInversionTrigger'
+        )} (${numOfActiveTextualInversions()})`}
+        menuButtonProps={{
+          width: '100%',
+          padding: '0 1rem',
+        }}
+      />
+      <IAIIconButton
+        icon={<MdArrowDownward />}
+        style={{
+          backgroundColor: addTIToNegative ? 'var(--btn-delete-image)' : '',
+        }}
+        tooltip={t('modelManager.addTIToNegative')}
+        aria-label={t('modelManager.addTIToNegative')}
+        onClick={() => dispatch(setAddTIToNegative(!addTIToNegative))}
+      />
+      <IAIIconButton
+        icon={<MdClear />}
+        tooltip={t('modelManager.clearTextualInversions')}
+        aria-label={t('modelManager.clearTextualInversions')}
+        onClick={() => dispatch(setClearTextualInversions())}
+      />
+    </Flex>
+  ) : (
+    <Box
+      background="var(--btn-base-color)"
+      padding={2}
+      textAlign="center"
+      borderRadius={4}
+      fontWeight="bold"
+    >
+      {t('modelManager.noTextualInversionTriggers')}
+    </Box>
+  );
+}

invokeai/frontend/src/features/parameters/components/PromptInput/NegativePromptInput.tsx

@@ -1,24 +1,43 @@
 import { FormControl, Textarea } from '@chakra-ui/react';
 import type { RootState } from 'app/store';
 import { useAppDispatch, useAppSelector } from 'app/storeHooks';
-import { setNegativePrompt } from 'features/parameters/store/generationSlice';
+import {
+  handlePromptCheckers,
+  setNegativePrompt,
+} from 'features/parameters/store/generationSlice';
 import { useTranslation } from 'react-i18next';
+import { ChangeEvent, useState } from 'react';
 
 const NegativePromptInput = () => {
   const negativePrompt = useAppSelector(
     (state: RootState) => state.generation.negativePrompt
   );
 
+  const [promptTimer, setPromptTimer] = useState<number | undefined>(undefined);
+
   const dispatch = useAppDispatch();
   const { t } = useTranslation();
 
+  const handleNegativeChangePrompt = (e: ChangeEvent<HTMLTextAreaElement>) => {
+    dispatch(setNegativePrompt(e.target.value));
+
+    // Debounce Prompt UI Checking
+    clearTimeout(promptTimer);
+    const newPromptTimer = window.setTimeout(() => {
+      dispatch(
+        handlePromptCheckers({ prompt: e.target.value, toNegative: true })
+      );
+    }, 500);
+    setPromptTimer(newPromptTimer);
+  };
+
   return (
     <FormControl>
       <Textarea
         id="negativePrompt"
         name="negativePrompt"
         value={negativePrompt}
-        onChange={(e) => dispatch(setNegativePrompt(e.target.value))}
+        onChange={handleNegativeChangePrompt}
         background="var(--prompt-bg-color)"
         placeholder={t('parameters.negativePrompts')}
         _placeholder={{ fontSize: '0.8rem' }}

invokeai/frontend/src/features/parameters/components/PromptInput/PromptInput.tsx

@@ -2,12 +2,13 @@ import { FormControl, Textarea } from '@chakra-ui/react';
 import { generateImage } from 'app/socketio/actions';
 import { RootState } from 'app/store';
 import { useAppDispatch, useAppSelector } from 'app/storeHooks';
-import { ChangeEvent, KeyboardEvent, useRef } from 'react';
+import { ChangeEvent, KeyboardEvent, useRef, useState } from 'react';
 
 import { createSelector } from '@reduxjs/toolkit';
 import { readinessSelector } from 'app/selectors/readinessSelector';
 import {
   GenerationState,
+  handlePromptCheckers,
   setPrompt,
 } from 'features/parameters/store/generationSlice';
 import { activeTabNameSelector } from 'features/ui/store/uiSelectors';
@@ -40,11 +41,21 @@ const PromptInput = () => {
   const { isReady } = useAppSelector(readinessSelector);
 
   const promptRef = useRef<HTMLTextAreaElement>(null);
+  const [promptTimer, setPromptTimer] = useState<number | undefined>(undefined);
 
   const { t } = useTranslation();
 
   const handleChangePrompt = (e: ChangeEvent<HTMLTextAreaElement>) => {
     dispatch(setPrompt(e.target.value));
+
+    // Debounce Prompt UI Checking
+    clearTimeout(promptTimer);
+    const newPromptTimer = window.setTimeout(() => {
+      dispatch(
+        handlePromptCheckers({ prompt: e.target.value, toNegative: false })
+      );
+    }, 500);
+    setPromptTimer(newPromptTimer);
   };
 
   useHotkeys(

invokeai/frontend/src/features/parameters/hooks/usePrompt.ts

@@ -3,7 +3,11 @@ import { getPromptAndNegative } from 'common/util/getPromptAndNegative';
 import * as InvokeAI from 'app/invokeai';
 import promptToString from 'common/util/promptToString';
 import { useAppDispatch } from 'app/storeHooks';
-import { setNegativePrompt, setPrompt } from '../store/generationSlice';
+import {
+  handlePromptCheckers,
+  setNegativePrompt,
+  setPrompt,
+} from '../store/generationSlice';
 
 // TECHDEBT: We have two metadata prompt formats and need to handle recalling either of them.
 // This hook provides a function to do that.
@@ -20,6 +24,10 @@ const useSetBothPrompts = () => {
 
     dispatch(setPrompt(prompt));
     dispatch(setNegativePrompt(negativePrompt));
+    dispatch(handlePromptCheckers({ prompt: prompt, toNegative: false }));
+    dispatch(
+      handlePromptCheckers({ prompt: negativePrompt, toNegative: true })
+    );
   };
 };
 

invokeai/frontend/src/features/parameters/store/generationSlice.ts

@@ -1,4 +1,4 @@
-import type { PayloadAction } from '@reduxjs/toolkit';
+import { PayloadAction } from '@reduxjs/toolkit';
 import { createSlice } from '@reduxjs/toolkit';
 import * as InvokeAI from 'app/invokeai';
 import { getPromptAndNegative } from 'common/util/getPromptAndNegative';
@@ -17,6 +17,12 @@ export interface GenerationState {
   perlin: number;
   prompt: string;
   negativePrompt: string;
+  lorasInUse: string[];
+  huggingFaceTextualInversionConcepts: string[];
+  localTextualInversionTriggers: string[];
+  textualInversionsInUse: string[];
+  negativeTextualInversionsInUse: string[];
+  addTIToNegative: boolean;
   sampler: string;
   seamBlur: number;
   seamless: boolean;
@@ -48,6 +54,12 @@ const initialGenerationState: GenerationState = {
   perlin: 0,
   prompt: '',
   negativePrompt: '',
+  lorasInUse: [],
+  huggingFaceTextualInversionConcepts: [],
+  localTextualInversionTriggers: [],
+  textualInversionsInUse: [],
+  negativeTextualInversionsInUse: [],
+  addTIToNegative: false,
   sampler: 'k_lms',
   seamBlur: 16,
   seamless: false,
@@ -71,12 +83,99 @@ const initialGenerationState: GenerationState = {
 
 const initialState: GenerationState = initialGenerationState;
 
+const loraExists = (state: GenerationState, lora: string) => {
+  const loraRegex = new RegExp(`withLora\\(${lora},?\\s*([^\\)]+)?\\)`);
+  if (state.prompt.match(loraRegex)) return true;
+  return false;
+};
+
+const getTIRegex = (textualInversion: string) => {
+  if (textualInversion.includes('<' || '>')) {
+    return new RegExp(`${textualInversion}`);
+  } else {
+    return new RegExp(`\\b${textualInversion}\\b`);
+  }
+};
+
+const textualInversionExists = (
+  state: GenerationState,
+  textualInversion: string
+) => {
+  const textualInversionRegex = getTIRegex(textualInversion);
+
+  if (!state.addTIToNegative) {
+    if (state.prompt.match(textualInversionRegex)) return true;
+  } else {
+    if (state.negativePrompt.match(textualInversionRegex)) return true;
+  }
+  return false;
+};
+
+const handleTypedTICheck = (
+  state: GenerationState,
+  newPrompt: string,
+  toNegative: boolean
+) => {
+  let textualInversionsInUse = !toNegative
+    ? [...state.textualInversionsInUse]
+    : [...state.negativeTextualInversionsInUse]; // Get Words In Prompt
+
+  const textualInversionRegex = /([\w<>!@%&*_-]+)/g; // Scan For Each Word
+
+  const textualInversionMatches = [
+    ...newPrompt.matchAll(textualInversionRegex),
+  ]; // Match All Words
+
+  if (textualInversionMatches.length > 0) {
+    textualInversionsInUse = []; // Reset Textual Inversions In Use
+
+    textualInversionMatches.forEach((textualInversionMatch) => {
+      const textualInversionName = textualInversionMatch[0];
+      if (
+        (!textualInversionsInUse.includes(textualInversionName) &&
+          state.localTextualInversionTriggers.includes(textualInversionName)) ||
+        state.huggingFaceTextualInversionConcepts.includes(textualInversionName)
+      ) {
+        textualInversionsInUse.push(textualInversionName); // Add Textual Inversions In Prompt
+      }
+    });
+  } else {
+    textualInversionsInUse = []; // If No Matches, Remove Textual Inversions In Use
+  }
+
+  if (!toNegative) {
+    state.textualInversionsInUse = textualInversionsInUse;
+  } else {
+    state.negativeTextualInversionsInUse = textualInversionsInUse;
+  }
+};
+
+const handleTypedLoraCheck = (state: GenerationState, newPrompt: string) => {
+  let lorasInUse = [...state.lorasInUse]; // Get Loras In Prompt
+
+  const loraRegex = /withLora\(([^\\)]+)\)/g; // Scan For Lora Syntax
+  const loraMatches = [...newPrompt.matchAll(loraRegex)]; // Match All Lora Syntaxes
+
+  if (loraMatches.length > 0) {
+    lorasInUse = []; // Reset Loras In Use
+    loraMatches.forEach((loraMatch) => {
+      const loraName = loraMatch[1].split(',')[0];
+      if (!lorasInUse.includes(loraName)) lorasInUse.push(loraName); // Add Loras In Prompt
+    });
+  } else {
+    lorasInUse = []; // If No Matches, Remove Loras In Use
+  }
+
+  state.lorasInUse = lorasInUse;
+};
+
 export const generationSlice = createSlice({
   name: 'generation',
   initialState,
   reducers: {
     setPrompt: (state, action: PayloadAction<string | InvokeAI.Prompt>) => {
       const newPrompt = action.payload;
+
       if (typeof newPrompt === 'string') {
         state.prompt = newPrompt;
       } else {
@@ -94,6 +193,136 @@ export const generationSlice = createSlice({
         state.negativePrompt = promptToString(newPrompt);
       }
     },
+    handlePromptCheckers: (
+      state,
+      action: PayloadAction<{
+        prompt: string | InvokeAI.Prompt;
+        toNegative: boolean;
+      }>
+    ) => {
+      const newPrompt = action.payload.prompt;
+
+      if (typeof newPrompt === 'string') {
+        if (!action.payload.toNegative) handleTypedLoraCheck(state, newPrompt);
+        handleTypedTICheck(state, newPrompt, action.payload.toNegative);
+      }
+    },
+    setLorasInUse: (state, action: PayloadAction<string>) => {
+      const newLora = action.payload;
+      const loras = [...state.lorasInUse];
+
+      if (loraExists(state, newLora)) {
+        const loraRegex = new RegExp(
+          `withLora\\(${newLora},?\\s*([^\\)]+)?\\)`,
+          'g'
+        );
+        const newPrompt = state.prompt.replaceAll(loraRegex, '');
+        state.prompt = newPrompt.trim();
+
+        if (loras.includes(newLora)) {
+          const newLoraIndex = loras.indexOf(newLora);
+          if (newLoraIndex > -1) loras.splice(newLoraIndex, 1);
+        }
+      } else {
+        state.prompt = `${state.prompt.trim()} withLora(${newLora},0.75)`;
+        if (!loras.includes(newLora)) loras.push(newLora);
+      }
+      state.lorasInUse = loras;
+    },
+    setClearLoras: (state) => {
+      const lorasInUse = [...state.lorasInUse];
+
+      lorasInUse.forEach((lora) => {
+        const loraRegex = new RegExp(
+          `withLora\\(${lora},?\\s*([^\\)]+)?\\)`,
+          'g'
+        );
+        const newPrompt = state.prompt.replaceAll(loraRegex, '');
+        state.prompt = newPrompt.trim();
+      });
+
+      state.lorasInUse = [];
+    },
+    setTextualInversionsInUse: (state, action: PayloadAction<string>) => {
+      const newTextualInversion = action.payload;
+
+      const textualInversions = [...state.textualInversionsInUse];
+      const negativeTextualInversions = [
+        ...state.negativeTextualInversionsInUse,
+      ];
+
+      if (textualInversionExists(state, newTextualInversion)) {
+        const textualInversionRegex = getTIRegex(newTextualInversion);
+
+        if (!state.addTIToNegative) {
+          const newPrompt = state.prompt.replace(textualInversionRegex, '');
+          state.prompt = newPrompt.trim();
+
+          const newTIIndex = textualInversions.indexOf(newTextualInversion);
+          if (newTIIndex > -1) textualInversions.splice(newTIIndex, 1);
+        } else {
+          const newPrompt = state.negativePrompt.replace(
+            textualInversionRegex,
+            ''
+          );
+          state.negativePrompt = newPrompt.trim();
+
+          const newTIIndex =
+            negativeTextualInversions.indexOf(newTextualInversion);
+          if (newTIIndex > -1) negativeTextualInversions.splice(newTIIndex, 1);
+        }
+      } else {
+        if (!state.addTIToNegative) {
+          state.prompt = `${state.prompt.trim()} ${newTextualInversion}`;
+          textualInversions.push(newTextualInversion);
+        } else {
+          state.negativePrompt = `${state.negativePrompt.trim()} ${newTextualInversion}`;
+          negativeTextualInversions.push(newTextualInversion);
+        }
+      }
+
+      state.textualInversionsInUse = textualInversions;
+      state.negativeTextualInversionsInUse = negativeTextualInversions;
+    },
+    setClearTextualInversions: (state) => {
+      const textualInversions = [...state.textualInversionsInUse];
+      const negativeTextualInversions = [
+        ...state.negativeTextualInversionsInUse,
+      ];
+
+      textualInversions.forEach((ti) => {
+        const textualInversionRegex = getTIRegex(ti);
+        const newPrompt = state.prompt.replace(textualInversionRegex, '');
+        state.prompt = newPrompt.trim();
+      });
+
+      negativeTextualInversions.forEach((ti) => {
+        const textualInversionRegex = getTIRegex(ti);
+        const newPrompt = state.negativePrompt.replace(
+          textualInversionRegex,
+          ''
+        );
+        state.negativePrompt = newPrompt.trim();
+      });
+
+      state.textualInversionsInUse = [];
+      state.negativeTextualInversionsInUse = [];
+    },
+    setAddTIToNegative: (state, action: PayloadAction<boolean>) => {
+      state.addTIToNegative = action.payload;
+    },
+    setLocalTextualInversionTriggers: (
+      state,
+      action: PayloadAction<string[]>
+    ) => {
+      state.localTextualInversionTriggers = action.payload;
+    },
+    setHuggingFaceTextualInversionConcepts: (
+      state,
+      action: PayloadAction<string[]>
+    ) => {
+      state.huggingFaceTextualInversionConcepts = action.payload;
+    },
     setIterations: (state, action: PayloadAction<number>) => {
       state.iterations = action.payload;
     },
@@ -374,6 +603,14 @@ export const {
   setPerlin,
   setPrompt,
   setNegativePrompt,
+  handlePromptCheckers,
+  setLorasInUse,
+  setClearLoras,
+  setHuggingFaceTextualInversionConcepts,
+  setLocalTextualInversionTriggers,
+  setTextualInversionsInUse,
+  setAddTIToNegative,
+  setClearTextualInversions,
   setSampler,
   setSeamBlur,
   setSeamless,

invokeai/frontend/src/features/system/components/SettingsModal/SettingsModal.tsx

@@ -31,6 +31,7 @@ import {
 } from 'features/system/store/systemSlice';
 import { uiSelector } from 'features/ui/store/uiSelectors';
 import {
+  setShouldShowHuggingFaceConcepts,
   setShouldUseCanvasBetaLayout,
   setShouldUseSliders,
 } from 'features/ui/store/uiSlice';
@@ -52,7 +53,11 @@ const selector = createSelector(
       enableImageDebugging,
     } = system;
 
-    const { shouldUseCanvasBetaLayout, shouldUseSliders } = ui;
+    const {
+      shouldUseCanvasBetaLayout,
+      shouldUseSliders,
+      shouldShowHuggingFaceConcepts,
+    } = ui;
 
     return {
       shouldDisplayInProgressType,
@@ -63,6 +68,7 @@ const selector = createSelector(
       enableImageDebugging,
       shouldUseCanvasBetaLayout,
       shouldUseSliders,
+      shouldShowHuggingFaceConcepts,
     };
   },
   {
@@ -107,6 +113,7 @@ const SettingsModal = ({ children }: SettingsModalProps) => {
     enableImageDebugging,
     shouldUseCanvasBetaLayout,
     shouldUseSliders,
+    shouldShowHuggingFaceConcepts,
   } = useAppSelector(selector);
 
   /**
@@ -206,6 +213,14 @@ const SettingsModal = ({ children }: SettingsModalProps) => {
                   dispatch(setShouldUseSliders(e.target.checked))
                 }
               />
+              <IAISwitch
+                styleClass="settings-modal-item"
+                label={t('settings.showHuggingFaceConcepts')}
+                isChecked={shouldShowHuggingFaceConcepts}
+                onChange={(e: ChangeEvent<HTMLInputElement>) =>
+                  dispatch(setShouldShowHuggingFaceConcepts(e.target.checked))
+                }
+              />
             </div>
 
             <div className="settings-modal-items">

invokeai/frontend/src/features/system/store/systemSlice.ts

@@ -51,6 +51,13 @@ export interface SystemState
   toastQueue: UseToastOptions[];
   searchFolder: string | null;
   foundModels: InvokeAI.FoundModel[] | null;
+  foundLoras: InvokeAI.FoundLora[] | null;
+  foundLocalTextualInversionTriggers:
+    | InvokeAI.FoundTextualInversionTriggers[]
+    | null;
+  foundHuggingFaceTextualInversionTriggers:
+    | InvokeAI.FoundTextualInversionTriggers[]
+    | null;
   openModel: string | null;
   cancelOptions: {
     cancelType: CancelType;
@@ -93,6 +100,9 @@ const initialSystemState: SystemState = {
   toastQueue: [],
   searchFolder: null,
   foundModels: null,
+  foundLoras: null,
+  foundLocalTextualInversionTriggers: null,
+  foundHuggingFaceTextualInversionTriggers: null,
   openModel: null,
   cancelOptions: {
     cancelType: 'immediate',
@@ -262,6 +272,24 @@ export const systemSlice = createSlice({
     ) => {
       state.foundModels = action.payload;
     },
+    setFoundLoras: (
+      state,
+      action: PayloadAction<InvokeAI.FoundLora[] | null>
+    ) => {
+      state.foundLoras = action.payload;
+    },
+    setFoundLocalTextualInversionTriggers: (
+      state,
+      action: PayloadAction<InvokeAI.FoundTextualInversionTriggers[] | null>
+    ) => {
+      state.foundLocalTextualInversionTriggers = action.payload;
+    },
+    setFoundHuggingFaceTextualInversionTriggers: (
+      state,
+      action: PayloadAction<InvokeAI.FoundTextualInversionTriggers[] | null>
+    ) => {
+      state.foundHuggingFaceTextualInversionTriggers = action.payload;
+    },
     setOpenModel: (state, action: PayloadAction<string | null>) => {
       state.openModel = action.payload;
     },
@@ -303,6 +331,9 @@ export const {
   setProcessingIndeterminateTask,
   setSearchFolder,
   setFoundModels,
+  setFoundLoras,
+  setFoundLocalTextualInversionTriggers,
+  setFoundHuggingFaceTextualInversionTriggers,
   setOpenModel,
   setCancelType,
   setCancelAfter,

invokeai/frontend/src/features/ui/components/ImageToImage/ImageToImagePanel.tsx

@@ -18,6 +18,7 @@ import PromptInput from 'features/parameters/components/PromptInput/PromptInput'
 import InvokeOptionsPanel from 'features/ui/components/InvokeParametersPanel';
 import { useTranslation } from 'react-i18next';
 import ImageToImageOptions from './ImageToImageOptions';
+import PromptExtras from 'features/parameters/components/PromptInput/Extras/PromptExtras';
 
 export default function ImageToImagePanel() {
   const { t } = useTranslation();
@@ -63,6 +64,7 @@ export default function ImageToImagePanel() {
       <Flex flexDir="column" rowGap="0.5rem">
         <PromptInput />
         <NegativePromptInput />
+        <PromptExtras />
       </Flex>
       <ProcessButtons />
       <MainSettings />

invokeai/frontend/src/features/ui/components/TextToImage/TextToImagePanel.tsx

@@ -17,6 +17,7 @@ import NegativePromptInput from 'features/parameters/components/PromptInput/Nega
 import PromptInput from 'features/parameters/components/PromptInput/PromptInput';
 import InvokeOptionsPanel from 'features/ui/components/InvokeParametersPanel';
 import { useTranslation } from 'react-i18next';
+import PromptExtras from 'features/parameters/components/PromptInput/Extras/PromptExtras';
 
 export default function TextToImagePanel() {
   const { t } = useTranslation();
@@ -62,6 +63,7 @@ export default function TextToImagePanel() {
       <Flex flexDir="column" rowGap="0.5rem">
         <PromptInput />
         <NegativePromptInput />
+        <PromptExtras />
       </Flex>
       <ProcessButtons />
       <MainSettings />

invokeai/frontend/src/features/ui/components/UnifiedCanvas/UnifiedCanvasPanel.tsx

@@ -17,6 +17,7 @@ import NegativePromptInput from 'features/parameters/components/PromptInput/Nega
 import PromptInput from 'features/parameters/components/PromptInput/PromptInput';
 import InvokeOptionsPanel from 'features/ui/components/InvokeParametersPanel';
 import { useTranslation } from 'react-i18next';
+import PromptExtras from 'features/parameters/components/PromptInput/Extras/PromptExtras';
 
 export default function UnifiedCanvasPanel() {
   const { t } = useTranslation();
@@ -73,6 +74,7 @@ export default function UnifiedCanvasPanel() {
       <Flex flexDir="column" rowGap="0.5rem">
         <PromptInput />
         <NegativePromptInput />
+        <PromptExtras />
       </Flex>
       <ProcessButtons />
       <MainSettings />

invokeai/frontend/src/features/ui/store/uiSlice.ts

@@ -15,6 +15,7 @@ const initialtabsState: UIState = {
   shouldUseCanvasBetaLayout: false,
   shouldShowExistingModelsInSearch: false,
   shouldUseSliders: false,
+  shouldShowHuggingFaceConcepts: false,
   addNewModelUIOption: null,
 };
 
@@ -70,6 +71,12 @@ export const uiSlice = createSlice({
     setShouldUseSliders: (state, action: PayloadAction<boolean>) => {
       state.shouldUseSliders = action.payload;
     },
+    setShouldShowHuggingFaceConcepts: (
+      state,
+      action: PayloadAction<boolean>
+    ) => {
+      state.shouldShowHuggingFaceConcepts = action.payload;
+    },
     setAddNewModelUIOption: (state, action: PayloadAction<AddNewModelType>) => {
       state.addNewModelUIOption = action.payload;
     },
@@ -88,6 +95,7 @@ export const {
   setShouldUseCanvasBetaLayout,
   setShouldShowExistingModelsInSearch,
   setShouldUseSliders,
+  setShouldShowHuggingFaceConcepts,
   setAddNewModelUIOption,
 } = uiSlice.actions;
 

invokeai/frontend/src/features/ui/store/uiTypes.ts

@@ -12,5 +12,6 @@ export interface UIState {
   shouldUseCanvasBetaLayout: boolean;
   shouldShowExistingModelsInSearch: boolean;
   shouldUseSliders: boolean;
+  shouldShowHuggingFaceConcepts: boolean;
   addNewModelUIOption: AddNewModelType;
 }

invokeai/frontend/src/styles/Themes/_Colors_Dark.scss

@@ -8,7 +8,6 @@
   --accent-color-bright: rgb(104, 60, 230);
   --accent-color-hover: var(--accent-color-bright);
 
-  // App Colors
   --root-bg-color: rgb(10, 10, 10);
   --background-color: rgb(26, 26, 32);
   --background-color-light: rgb(40, 44, 48);
@@ -119,6 +118,7 @@
   --context-menu-bg-color: rgb(46, 48, 58);
   --context-menu-box-shadow: none;
   --context-menu-bg-color-hover: rgb(30, 32, 42);
+  --context-menu-active-item: var(--accent-color-bright);
 
   // Shadows
   --floating-button-drop-shadow-color: var(--accent-color);

invokeai/frontend/src/styles/Themes/_Colors_Green.scss

@@ -117,6 +117,7 @@
   --context-menu-bg-color: rgb(46, 48, 58);
   --context-menu-box-shadow: none;
   --context-menu-bg-color-hover: rgb(30, 32, 42);
+  --context-menu-active-item: var(--accent-color-bright);
 
   // Shadows
   --floating-button-drop-shadow-color: var(--accent-color);

invokeai/frontend/src/styles/Themes/_Colors_Light.scss

@@ -114,6 +114,7 @@
   --context-menu-box-shadow: 0px 10px 38px -10px rgba(22, 23, 24, 0.35),
     0px 10px 20px -15px rgba(22, 23, 24, 0.2);
   --context-menu-bg-color-hover: var(--background-color-secondary);
+  --context-menu-active-item: rgb(0, 0, 0);
 
   // Shadows
   --floating-button-drop-shadow-color: rgba(0, 0, 0, 0.7);

invokeai/frontend/yarn.lock

@@ -56,26 +56,26 @@
     "@babel/helper-validator-identifier" "^7.19.1"
     to-fast-properties "^2.0.0"
 
-"@chakra-ui/accordion@2.1.9":
-  version "2.1.9"
-  resolved "https://registry.yarnpkg.com/@chakra-ui/accordion/-/accordion-2.1.9.tgz#20fa86d94dc034251df2f7c8595ae4dd541a29d9"
-  integrity sha512-a9CKIAUHezc0f5FR/SQ4GVxnWuIb2HbDTxTEKTp58w/J9pecIbJaNrJ5TUZ0MVbDU9jkgO9RsZ29jkja8PomAw==
+"@chakra-ui/accordion@2.1.11":
+  version "2.1.11"
+  resolved "https://registry.yarnpkg.com/@chakra-ui/accordion/-/accordion-2.1.11.tgz#c6df0100c543645d0631df3aefde2ea2b8ed6313"
+  integrity sha512-mfVPmqETp9pyRDHJ33AdF19oHv/LyxVzQJtlxUByuvs8Cj9QQZ2LQLg5kejm+b3mj03A7A6yfbuo3RNaI4Bhsg==
   dependencies:
-    "@chakra-ui/descendant" "3.0.13"
+    "@chakra-ui/descendant" "3.0.14"
     "@chakra-ui/icon" "3.0.16"
-    "@chakra-ui/react-context" "2.0.7"
+    "@chakra-ui/react-context" "2.0.8"
     "@chakra-ui/react-use-controllable-state" "2.0.8"
     "@chakra-ui/react-use-merge-refs" "2.0.7"
     "@chakra-ui/shared-utils" "2.0.5"
-    "@chakra-ui/transition" "2.0.15"
+    "@chakra-ui/transition" "2.0.16"
 
-"@chakra-ui/alert@2.0.17":
-  version "2.0.17"
-  resolved "https://registry.yarnpkg.com/@chakra-ui/alert/-/alert-2.0.17.tgz#b129732ec308db6a6a1afa7c06a6595ad853c967"
-  integrity sha512-0Y5vw+HkeXpwbL1roVpSSNM6luMRmUbwduUSHEA4OnX1ismvsDb1ZBfpi4Vxp6w8euJ2Uj6df3krbd5tbCP6tg==
+"@chakra-ui/alert@2.1.0":
+  version "2.1.0"
+  resolved "https://registry.yarnpkg.com/@chakra-ui/alert/-/alert-2.1.0.tgz#7a234ac6426231b39243088648455cbcf1cbdf24"
+  integrity sha512-OcfHwoXI5VrmM+tHJTHT62Bx6TfyfCxSa0PWUOueJzSyhlUOKBND5we6UtrOB7D0jwX45qKKEDJOLG5yCG21jQ==
   dependencies:
     "@chakra-ui/icon" "3.0.16"
-    "@chakra-ui/react-context" "2.0.7"
+    "@chakra-ui/react-context" "2.0.8"
     "@chakra-ui/shared-utils" "2.0.5"
     "@chakra-ui/spinner" "2.0.13"
 
@@ -84,23 +84,23 @@
   resolved "https://registry.yarnpkg.com/@chakra-ui/anatomy/-/anatomy-2.1.2.tgz#ea66b1841e7195da08ddc862daaa3f3e56e565f5"
   integrity sha512-pKfOS/mztc4sUXHNc8ypJ1gPWSolWT770jrgVRfolVbYlki8y5Y+As996zMF6k5lewTu6j9DQequ7Cc9a69IVQ==
 
-"@chakra-ui/avatar@2.2.5":
-  version "2.2.5"
-  resolved "https://registry.yarnpkg.com/@chakra-ui/avatar/-/avatar-2.2.5.tgz#50eb7cc5a172d394b301fa0abd5f607b7f5d3563"
-  integrity sha512-TEHXuGE79+fEn61qJ7J/A0Ec+WjyNwobrDTATcLg9Zx2/WEMmZNfrWIAlI5ANQAwVbdSWeGVbyoLAK5mbcrE0A==
+"@chakra-ui/avatar@2.2.8":
+  version "2.2.8"
+  resolved "https://registry.yarnpkg.com/@chakra-ui/avatar/-/avatar-2.2.8.tgz#a6e16accb2bb9c879f197090ccc9df1ff42992a6"
+  integrity sha512-uBs9PMrqyK111tPIYIKnOM4n3mwgKqGpvYmtwBnnbQLTNLg4gtiWWVbpTuNMpyu1av0xQYomjUt8Doed8w6p8g==
   dependencies:
     "@chakra-ui/image" "2.0.15"
     "@chakra-ui/react-children-utils" "2.0.6"
-    "@chakra-ui/react-context" "2.0.7"
+    "@chakra-ui/react-context" "2.0.8"
     "@chakra-ui/shared-utils" "2.0.5"
 
-"@chakra-ui/breadcrumb@2.1.4":
-  version "2.1.4"
-  resolved "https://registry.yarnpkg.com/@chakra-ui/breadcrumb/-/breadcrumb-2.1.4.tgz#0d249dc2a92639bd2bf46d097dd5445112bd2367"
-  integrity sha512-vyBx5TAxPnHhb0b8nyRGfqyjleD//9mySFhk96c9GL+T6YDO4swHw5y/kvDv3Ngc/iRwJ9hdI49PZKwPxLqsEg==
+"@chakra-ui/breadcrumb@2.1.5":
+  version "2.1.5"
+  resolved "https://registry.yarnpkg.com/@chakra-ui/breadcrumb/-/breadcrumb-2.1.5.tgz#a43b22cc8005291a615696a8c88efc37064562f3"
+  integrity sha512-p3eQQrHQBkRB69xOmNyBJqEdfCrMt+e0eOH+Pm/DjFWfIVIbnIaFbmDCeWClqlLa21Ypc6h1hR9jEmvg8kmOog==
   dependencies:
     "@chakra-ui/react-children-utils" "2.0.6"
-    "@chakra-ui/react-context" "2.0.7"
+    "@chakra-ui/react-context" "2.0.8"
     "@chakra-ui/shared-utils" "2.0.5"
 
 "@chakra-ui/breakpoint-utils@2.0.8":
@@ -110,12 +110,12 @@
   dependencies:
     "@chakra-ui/shared-utils" "2.0.5"
 
-"@chakra-ui/button@2.0.16":
-  version "2.0.16"
-  resolved "https://registry.yarnpkg.com/@chakra-ui/button/-/button-2.0.16.tgz#ff315b57ee47c3511a6507fcfb6f00bb93e2ac7d"
-  integrity sha512-NjuTKa7gNhnGSUutKuTc8HoAOe9WWIigpciBG7yj3ok67kg8bXtSzPyQFZlgTY6XGdAckWTT+Do4tvhwa5LA+g==
+"@chakra-ui/button@2.0.18":
+  version "2.0.18"
+  resolved "https://registry.yarnpkg.com/@chakra-ui/button/-/button-2.0.18.tgz#c13d2e404e22a9873ba5373fde494bedafe32fdd"
+  integrity sha512-E3c99+lOm6ou4nQVOTLkG+IdOPMjsQK+Qe7VyP8A/xeAMFONuibrWPRPpprr4ZkB4kEoLMfNuyH2+aEza3ScUA==
   dependencies:
-    "@chakra-ui/react-context" "2.0.7"
+    "@chakra-ui/react-context" "2.0.8"
     "@chakra-ui/react-use-merge-refs" "2.0.7"
     "@chakra-ui/shared-utils" "2.0.5"
     "@chakra-ui/spinner" "2.0.13"
@@ -127,13 +127,13 @@
   dependencies:
     "@chakra-ui/shared-utils" "2.0.5"
 
-"@chakra-ui/checkbox@2.2.10":
-  version "2.2.10"
-  resolved "https://registry.yarnpkg.com/@chakra-ui/checkbox/-/checkbox-2.2.10.tgz#e4f773e7d2464f1d6e9d18dd88b679290cb33171"
-  integrity sha512-vzxEjw99qj7loxAdP1WuHNt4EAvj/t6cc8oxyOB2mEvkAzhxI34rLR+3zWDuHWsmhyUO+XEDh4FiWdR+DK5Siw==
+"@chakra-ui/checkbox@2.2.14":
+  version "2.2.14"
+  resolved "https://registry.yarnpkg.com/@chakra-ui/checkbox/-/checkbox-2.2.14.tgz#902acc99a9a80c1c304788a230cf36f8116e8260"
+  integrity sha512-uqo6lFWLqYBujPglrvRhTAErtuIXpmdpc5w0W4bjK7kyvLhxOpUh1hlDb2WoqlNpfRn/OaNeF6VinPnf9BJL8w==
   dependencies:
-    "@chakra-ui/form-control" "2.0.17"
-    "@chakra-ui/react-context" "2.0.7"
+    "@chakra-ui/form-control" "2.0.18"
+    "@chakra-ui/react-context" "2.0.8"
     "@chakra-ui/react-types" "2.0.7"
     "@chakra-ui/react-use-callback-ref" "2.0.7"
     "@chakra-ui/react-use-controllable-state" "2.0.8"
@@ -142,7 +142,7 @@
     "@chakra-ui/react-use-update-effect" "2.0.7"
     "@chakra-ui/shared-utils" "2.0.5"
     "@chakra-ui/visually-hidden" "2.0.15"
-    "@zag-js/focus-visible" "0.2.1"
+    "@zag-js/focus-visible" "0.2.2"
 
 "@chakra-ui/clickable@2.0.14":
   version "2.0.14"
@@ -180,17 +180,17 @@
     "@chakra-ui/react-use-callback-ref" "2.0.7"
     "@chakra-ui/shared-utils" "2.0.5"
 
-"@chakra-ui/css-reset@2.0.12":
-  version "2.0.12"
-  resolved "https://registry.yarnpkg.com/@chakra-ui/css-reset/-/css-reset-2.0.12.tgz#6eebcbe9e971facd215e174e063ace29f647a045"
-  integrity sha512-Q5OYIMvqTl2vZ947kIYxcS5DhQXeStB84BzzBd6C10wOx1gFUu9pL+jLpOnHR3hhpWRMdX5o7eT+gMJWIYUZ0Q==
+"@chakra-ui/css-reset@2.1.1":
+  version "2.1.1"
+  resolved "https://registry.yarnpkg.com/@chakra-ui/css-reset/-/css-reset-2.1.1.tgz#c61f3d2103c13e62a86fd2d359682092e961852c"
+  integrity sha512-jwEOfIAWmQsnChHQTW/eRE+dfE4MjmhvSvoUug5nkV1pI7veC/20noFlIZxzi82EbiQI8Fs0+Jnusgxr2yaOHA==
 
-"@chakra-ui/descendant@3.0.13":
-  version "3.0.13"
-  resolved "https://registry.yarnpkg.com/@chakra-ui/descendant/-/descendant-3.0.13.tgz#e883a2233ee07fe1ae6c014567824c0f79df11cf"
-  integrity sha512-9nzxZVxUSMc4xPL5fSaRkEOQjDQWUGjGvrZI7VzWk9eq63cojOtIxtWMSW383G9148PzWJjJYt30Eud5tdZzlg==
+"@chakra-ui/descendant@3.0.14":
+  version "3.0.14"
+  resolved "https://registry.yarnpkg.com/@chakra-ui/descendant/-/descendant-3.0.14.tgz#fe8bac3f0e1ffe562e3e73eac393dbf222d57e13"
+  integrity sha512-+Ahvp9H4HMpfScIv9w1vaecGz7qWAaK1YFHHolz/SIsGLaLGlbdp+5UNabQC7L6TUnzzJDQDxzwif78rTD7ang==
   dependencies:
-    "@chakra-ui/react-context" "2.0.7"
+    "@chakra-ui/react-context" "2.0.8"
     "@chakra-ui/react-use-merge-refs" "2.0.7"
 
 "@chakra-ui/dom-utils@2.0.6":
@@ -198,12 +198,12 @@
   resolved "https://registry.yarnpkg.com/@chakra-ui/dom-utils/-/dom-utils-2.0.6.tgz#68f49f3b4a0bdebd5e416d6fd2c012c9ad64b76a"
   integrity sha512-PVtDkPrDD5b8aoL6Atg7SLjkwhWb7BwMcLOF1L449L3nZN+DAO3nyAh6iUhZVJyunELj9d0r65CDlnMREyJZmA==
 
-"@chakra-ui/editable@2.0.19":
-  version "2.0.19"
-  resolved "https://registry.yarnpkg.com/@chakra-ui/editable/-/editable-2.0.19.tgz#1af2fe3c215111f61f7872fb5f599f4d8da24e7d"
-  integrity sha512-YxRJsJ2JQd42zfPBgTKzIhg1HugT+gfQz1ZosmUN+IZT9YZXL2yodHTUz6Lee04Vc/CdEqgBFLuREXEUNBfGtA==
+"@chakra-ui/editable@2.0.21":
+  version "2.0.21"
+  resolved "https://registry.yarnpkg.com/@chakra-ui/editable/-/editable-2.0.21.tgz#bc74510470d6d455844438e540851896d3879132"
+  integrity sha512-oYuXbHnggxSYJN7P9Pn0Scs9tPC91no4z1y58Oe+ILoJKZ+bFAEHtL7FEISDNJxw++MEukeFu7GU1hVqmdLsKQ==
   dependencies:
-    "@chakra-ui/react-context" "2.0.7"
+    "@chakra-ui/react-context" "2.0.8"
     "@chakra-ui/react-types" "2.0.7"
     "@chakra-ui/react-use-callback-ref" "2.0.7"
     "@chakra-ui/react-use-controllable-state" "2.0.8"
@@ -226,13 +226,13 @@
     "@chakra-ui/dom-utils" "2.0.6"
     react-focus-lock "^2.9.2"
 
-"@chakra-ui/form-control@2.0.17":
-  version "2.0.17"
-  resolved "https://registry.yarnpkg.com/@chakra-ui/form-control/-/form-control-2.0.17.tgz#2f710325e77ce35067337616d440f903b137bdd5"
-  integrity sha512-34ptCaJ2LNvQNOlB6MAKsmH1AkT1xo7E+3Vw10Urr81yTOjDTM/iU6vG3JKPfRDMyXeowPjXmutlnuk72SSjRg==
+"@chakra-ui/form-control@2.0.18":
+  version "2.0.18"
+  resolved "https://registry.yarnpkg.com/@chakra-ui/form-control/-/form-control-2.0.18.tgz#1923f293afde70b2b07ca731d98fef3660098c56"
+  integrity sha512-I0a0jG01IAtRPccOXSNugyRdUAe8Dy40ctqedZvznMweOXzbMCF1m+sHPLdWeWC/VI13VoAispdPY0/zHOdjsQ==
   dependencies:
     "@chakra-ui/icon" "3.0.16"
-    "@chakra-ui/react-context" "2.0.7"
+    "@chakra-ui/react-context" "2.0.8"
     "@chakra-ui/react-types" "2.0.7"
     "@chakra-ui/react-use-merge-refs" "2.0.7"
     "@chakra-ui/shared-utils" "2.0.5"
@@ -254,10 +254,10 @@
   dependencies:
     "@chakra-ui/shared-utils" "2.0.5"
 
-"@chakra-ui/icons@^2.0.17":
-  version "2.0.17"
-  resolved "https://registry.yarnpkg.com/@chakra-ui/icons/-/icons-2.0.17.tgz#625a46d169707aad36d65c04a4626a422f92e5ae"
-  integrity sha512-HMJP0WrJgAmFR9+Xh/CBH0nVnGMsJ4ZC8MK6tMgxPKd9/muvn0I4hsicHqdPlLpmB0TlxlhkBAKaVMtOdz6F0w==
+"@chakra-ui/icons@^2.0.18":
+  version "2.0.18"
+  resolved "https://registry.yarnpkg.com/@chakra-ui/icons/-/icons-2.0.18.tgz#6f859d2e0d8f31fea9cb2e6507d65eb65cb95cf5"
+  integrity sha512-E/+DF/jw7kdN4/XxCZRnr4FdMXhkl50Q34MVwN9rADWMwPK9uSZPGyC7HOx6rilo7q4bFjYDH3yRj9g+VfbVkg==
   dependencies:
     "@chakra-ui/icon" "3.0.16"
 
@@ -269,27 +269,27 @@
     "@chakra-ui/react-use-safe-layout-effect" "2.0.5"
     "@chakra-ui/shared-utils" "2.0.5"
 
-"@chakra-ui/input@2.0.20":
-  version "2.0.20"
-  resolved "https://registry.yarnpkg.com/@chakra-ui/input/-/input-2.0.20.tgz#8db3ec46b52be901c94599b3659a9003bdb2dd07"
-  integrity sha512-ypmsy4n4uNBVgn6Gd24Zrpi+qRf/T9WEzWkysuYC9Qfxo+i7yuf3snp7XmBy8KSGVSiXE11eO8ZN5oCg6Xg0jg==
+"@chakra-ui/input@2.0.21":
+  version "2.0.21"
+  resolved "https://registry.yarnpkg.com/@chakra-ui/input/-/input-2.0.21.tgz#a7e55ea6fa32ae39c0f6ec44ca2189933fda9eb5"
+  integrity sha512-AIWjjg6MgcOtlvKmVoZfPPfgF+sBSWL3Zq2HSCAMvS6h7jfxz/Xv0UTFGPk5F4Wt0YHT7qMySg0Jsm0b78HZJg==
   dependencies:
-    "@chakra-ui/form-control" "2.0.17"
+    "@chakra-ui/form-control" "2.0.18"
     "@chakra-ui/object-utils" "2.0.8"
     "@chakra-ui/react-children-utils" "2.0.6"
-    "@chakra-ui/react-context" "2.0.7"
+    "@chakra-ui/react-context" "2.0.8"
     "@chakra-ui/shared-utils" "2.0.5"
 
-"@chakra-ui/layout@2.1.16":
-  version "2.1.16"
-  resolved "https://registry.yarnpkg.com/@chakra-ui/layout/-/layout-2.1.16.tgz#9d90f25cf9f0537d19cd36a417f7ddc1461e8591"
-  integrity sha512-QFS3feozIGsvB0H74lUocev55aRF26eNrdmhfJifwikZAiq+zzZAMdBdNU9UJhHClnMOU8/iGZ0MF7ti4zQS1A==
+"@chakra-ui/layout@2.1.18":
+  version "2.1.18"
+  resolved "https://registry.yarnpkg.com/@chakra-ui/layout/-/layout-2.1.18.tgz#f5dba687dfced9145d495f3a21edb5672df6bb73"
+  integrity sha512-F4Gh2e+DGdaWdWT5NZduIFD9NM7Bnuh8sXARFHWPvIu7yvAwZ3ddqC9GK4F3qUngdmkJxDLWQqRSwSh96Lxbhw==
   dependencies:
     "@chakra-ui/breakpoint-utils" "2.0.8"
     "@chakra-ui/icon" "3.0.16"
     "@chakra-ui/object-utils" "2.0.8"
     "@chakra-ui/react-children-utils" "2.0.6"
-    "@chakra-ui/react-context" "2.0.7"
+    "@chakra-ui/react-context" "2.0.8"
     "@chakra-ui/shared-utils" "2.0.5"
 
 "@chakra-ui/lazy-utils@2.0.5":
@@ -311,17 +311,17 @@
     "@chakra-ui/react-env" "3.0.0"
     "@chakra-ui/shared-utils" "2.0.5"
 
-"@chakra-ui/menu@2.1.9":
-  version "2.1.9"
-  resolved "https://registry.yarnpkg.com/@chakra-ui/menu/-/menu-2.1.9.tgz#2f3239a9b2855fd77fc317d9e6b904c1ad50d7c6"
-  integrity sha512-ue5nD4QJcl3H3UwN0zZNJmH89XUebnvEdW6THAUL41hDjJ0J/Fjpg9Sgzwug2aBbBXBNbVMsUuhcCj6x91d+IQ==
+"@chakra-ui/menu@2.1.12":
+  version "2.1.12"
+  resolved "https://registry.yarnpkg.com/@chakra-ui/menu/-/menu-2.1.12.tgz#ab83b7a5165bd31a6c68328d7f65a79e3412c48d"
+  integrity sha512-ylNK1VJlr/3/EGg9dLPZ87cBJJjeiYXeU/gOAphsKXMnByrXWhbp4YVnyyyha2KZ0zEw0aPU4nCZ+A69aT9wrg==
   dependencies:
     "@chakra-ui/clickable" "2.0.14"
-    "@chakra-ui/descendant" "3.0.13"
+    "@chakra-ui/descendant" "3.0.14"
     "@chakra-ui/lazy-utils" "2.0.5"
     "@chakra-ui/popper" "3.0.13"
     "@chakra-ui/react-children-utils" "2.0.6"
-    "@chakra-ui/react-context" "2.0.7"
+    "@chakra-ui/react-context" "2.0.8"
     "@chakra-ui/react-use-animation-state" "2.0.8"
     "@chakra-ui/react-use-controllable-state" "2.0.8"
     "@chakra-ui/react-use-disclosure" "2.0.8"
@@ -330,33 +330,33 @@
     "@chakra-ui/react-use-outside-click" "2.0.7"
     "@chakra-ui/react-use-update-effect" "2.0.7"
     "@chakra-ui/shared-utils" "2.0.5"
-    "@chakra-ui/transition" "2.0.15"
+    "@chakra-ui/transition" "2.0.16"
 
-"@chakra-ui/modal@2.2.9":
-  version "2.2.9"
-  resolved "https://registry.yarnpkg.com/@chakra-ui/modal/-/modal-2.2.9.tgz#aad65a2c60aa974e023f8b3facc0e79eb742e006"
-  integrity sha512-nTfNp7XsVwn5+xJOtstoFA8j0kq/9sJj7KesyYzjEDaMKvCZvIOntRYowoydho43jb4+YC7ebKhp0KOIINS0gg==
+"@chakra-ui/modal@2.2.11":
+  version "2.2.11"
+  resolved "https://registry.yarnpkg.com/@chakra-ui/modal/-/modal-2.2.11.tgz#8a964288759f3d681e23bfc3a837a3e2c7523f8e"
+  integrity sha512-2J0ZUV5tEzkPiawdkgPz6bmex7NXAde1VXooMwdvK+vuT8PV3U61yorTJOZVLdw7TjjI1Yo94mzsp6UwBud43Q==
   dependencies:
     "@chakra-ui/close-button" "2.0.17"
     "@chakra-ui/focus-lock" "2.0.16"
-    "@chakra-ui/portal" "2.0.15"
-    "@chakra-ui/react-context" "2.0.7"
+    "@chakra-ui/portal" "2.0.16"
+    "@chakra-ui/react-context" "2.0.8"
     "@chakra-ui/react-types" "2.0.7"
     "@chakra-ui/react-use-merge-refs" "2.0.7"
     "@chakra-ui/shared-utils" "2.0.5"
-    "@chakra-ui/transition" "2.0.15"
+    "@chakra-ui/transition" "2.0.16"
     aria-hidden "^1.2.2"
     react-remove-scroll "^2.5.5"
 
-"@chakra-ui/number-input@2.0.18":
-  version "2.0.18"
-  resolved "https://registry.yarnpkg.com/@chakra-ui/number-input/-/number-input-2.0.18.tgz#072a00ef869ebafa4960cfdee8caae8208864289"
-  integrity sha512-cPkyAFFHHzeFBselrT1BtjlzMkJ6TKrTDUnHFlzqXy6aqeXuhrjFhMfXucjedSpOqedsP9ZbKFTdIAhu9DdL/A==
+"@chakra-ui/number-input@2.0.19":
+  version "2.0.19"
+  resolved "https://registry.yarnpkg.com/@chakra-ui/number-input/-/number-input-2.0.19.tgz#82d4522036904c04d07e7050822fc522f9b32233"
+  integrity sha512-HDaITvtMEqOauOrCPsARDxKD9PSHmhWywpcyCSOX0lMe4xx2aaGhU0QQFhsJsykj8Er6pytMv6t0KZksdDv3YA==
   dependencies:
     "@chakra-ui/counter" "2.0.14"
-    "@chakra-ui/form-control" "2.0.17"
+    "@chakra-ui/form-control" "2.0.18"
     "@chakra-ui/icon" "3.0.16"
-    "@chakra-ui/react-context" "2.0.7"
+    "@chakra-ui/react-context" "2.0.8"
     "@chakra-ui/react-types" "2.0.7"
     "@chakra-ui/react-use-callback-ref" "2.0.7"
     "@chakra-ui/react-use-event-listener" "2.0.7"
@@ -376,27 +376,27 @@
   resolved "https://registry.yarnpkg.com/@chakra-ui/object-utils/-/object-utils-2.0.8.tgz#307f927f6434f99feb32ba92bdf451a6b59a6199"
   integrity sha512-2upjT2JgRuiupdrtBWklKBS6tqeGMA77Nh6Q0JaoQuH/8yq+15CGckqn3IUWkWoGI0Fg3bK9LDlbbD+9DLw95Q==
 
-"@chakra-ui/pin-input@2.0.19":
-  version "2.0.19"
-  resolved "https://registry.yarnpkg.com/@chakra-ui/pin-input/-/pin-input-2.0.19.tgz#f9b196174f0518feec5c1ee3fcaf2134c301148a"
-  integrity sha512-6O7s4vWz4cqQ6zvMov9sYj6ZqWAsTxR/MNGe3DNgu1zWQg8veNCYtj1rNGhNS3eZNUMAa8uM2dXIphGTP53Xow==
+"@chakra-ui/pin-input@2.0.20":
+  version "2.0.20"
+  resolved "https://registry.yarnpkg.com/@chakra-ui/pin-input/-/pin-input-2.0.20.tgz#5bf115bf4282b69fc6532a9c542cbf41f815d200"
+  integrity sha512-IHVmerrtHN8F+jRB3W1HnMir1S1TUCWhI7qDInxqPtoRffHt6mzZgLZ0izx8p1fD4HkW4c1d4/ZLEz9uH9bBRg==
   dependencies:
-    "@chakra-ui/descendant" "3.0.13"
+    "@chakra-ui/descendant" "3.0.14"
     "@chakra-ui/react-children-utils" "2.0.6"
-    "@chakra-ui/react-context" "2.0.7"
+    "@chakra-ui/react-context" "2.0.8"
     "@chakra-ui/react-use-controllable-state" "2.0.8"
     "@chakra-ui/react-use-merge-refs" "2.0.7"
     "@chakra-ui/shared-utils" "2.0.5"
 
-"@chakra-ui/popover@2.1.8":
-  version "2.1.8"
-  resolved "https://registry.yarnpkg.com/@chakra-ui/popover/-/popover-2.1.8.tgz#e906ce0533693d735b6e13a3a6ffe16d8e0a9ab4"
-  integrity sha512-ob7fAz+WWmXIq7iGHVB3wDKzZTj+T+noYBT/U1Q+jIf+jMr2WOpJLTfb0HTZcfhvn4EBFlfBg7Wk5qbXNaOn7g==
+"@chakra-ui/popover@2.1.9":
+  version "2.1.9"
+  resolved "https://registry.yarnpkg.com/@chakra-ui/popover/-/popover-2.1.9.tgz#890cc0dfc5022757715ccf772ec194e7a409275f"
+  integrity sha512-OMJ12VVs9N32tFaZSOqikkKPtwAVwXYsES/D1pff/amBrE3ngCrpxJSIp4uvTdORfIYDojJqrR52ZplDKS9hRQ==
   dependencies:
     "@chakra-ui/close-button" "2.0.17"
     "@chakra-ui/lazy-utils" "2.0.5"
     "@chakra-ui/popper" "3.0.13"
-    "@chakra-ui/react-context" "2.0.7"
+    "@chakra-ui/react-context" "2.0.8"
     "@chakra-ui/react-types" "2.0.7"
     "@chakra-ui/react-use-animation-state" "2.0.8"
     "@chakra-ui/react-use-disclosure" "2.0.8"
@@ -414,53 +414,53 @@
     "@chakra-ui/react-use-merge-refs" "2.0.7"
     "@popperjs/core" "^2.9.3"
 
-"@chakra-ui/portal@2.0.15":
-  version "2.0.15"
-  resolved "https://registry.yarnpkg.com/@chakra-ui/portal/-/portal-2.0.15.tgz#21e1f97c4407fc15df8c365cb5cf799dac73ce41"
-  integrity sha512-z8v7K3j1/nMuBzp2+wRIIw7s/eipVtnXLdjK5yqbMxMRa44E8Mu5VNJLz3aQFLHXEUST+ifqrjImQeli9do6LQ==
+"@chakra-ui/portal@2.0.16":
+  version "2.0.16"
+  resolved "https://registry.yarnpkg.com/@chakra-ui/portal/-/portal-2.0.16.tgz#e5ce3f9d9e559f17a95276e0c006d0e9b7703442"
+  integrity sha512-bVID0qbQ0l4xq38LdqAN4EKD4/uFkDnXzFwOlviC9sl0dNhzICDb1ltuH/Adl1d2HTMqyN60O3GO58eHy7plnQ==
   dependencies:
-    "@chakra-ui/react-context" "2.0.7"
+    "@chakra-ui/react-context" "2.0.8"
     "@chakra-ui/react-use-safe-layout-effect" "2.0.5"
 
-"@chakra-ui/progress@2.1.5":
-  version "2.1.5"
-  resolved "https://registry.yarnpkg.com/@chakra-ui/progress/-/progress-2.1.5.tgz#eb6a47adf2bff93971262d163461d390782a04ff"
-  integrity sha512-jj5Vp4lxUchuwp4RPCepM0yAyKi344bgsOd3Apd+ldxclDcewPc82fbwDu7g/Xv27LqJkT+7E/SlQy04wGrk0g==
+"@chakra-ui/progress@2.1.6":
+  version "2.1.6"
+  resolved "https://registry.yarnpkg.com/@chakra-ui/progress/-/progress-2.1.6.tgz#398db20440979c37adb0a34821f805ae3471873b"
+  integrity sha512-hHh5Ysv4z6bK+j2GJbi/FT9CVyto2PtNUNwBmr3oNMVsoOUMoRjczfXvvYqp0EHr9PCpxqrq7sRwgQXUzhbDSw==
   dependencies:
-    "@chakra-ui/react-context" "2.0.7"
+    "@chakra-ui/react-context" "2.0.8"
 
-"@chakra-ui/provider@2.1.2":
-  version "2.1.2"
-  resolved "https://registry.yarnpkg.com/@chakra-ui/provider/-/provider-2.1.2.tgz#b025cb718826b003b3c9535b6961e8f3be70ebd5"
-  integrity sha512-4lLlz8QuJv00BhfyKzWpzfoti9MDOdJ/MqXixJV/EZ02RMBOdE9qy9bSz/WckPC2MVhtRUuwMkxH+0QY21PXuw==
+"@chakra-ui/provider@2.2.2":
+  version "2.2.2"
+  resolved "https://registry.yarnpkg.com/@chakra-ui/provider/-/provider-2.2.2.tgz#a798d1c243f33e00c85763834a7350e0d1c643ad"
+  integrity sha512-UVwnIDnAWq1aKroN5AF+OpNpUqLVeIUk7tKvX3z4CY9FsPFFi6LTEhRHdhpwaU1Tau3Tf9agEu5URegpY7S8BA==
   dependencies:
-    "@chakra-ui/css-reset" "2.0.12"
-    "@chakra-ui/portal" "2.0.15"
+    "@chakra-ui/css-reset" "2.1.1"
+    "@chakra-ui/portal" "2.0.16"
     "@chakra-ui/react-env" "3.0.0"
-    "@chakra-ui/system" "2.5.1"
+    "@chakra-ui/system" "2.5.5"
     "@chakra-ui/utils" "2.0.15"
 
-"@chakra-ui/radio@2.0.19":
-  version "2.0.19"
-  resolved "https://registry.yarnpkg.com/@chakra-ui/radio/-/radio-2.0.19.tgz#8d5c02eae8eddbced4476b1b50921ade62f0a744"
-  integrity sha512-PlJiV59eGSmeKP4v/4+ccQUWGRd0cjPKkj/p3L+UbOf8pl9dWm8y9kIeL5TYbghQSDv0nzkrH4+yMnnDTZjdMQ==
+"@chakra-ui/radio@2.0.22":
+  version "2.0.22"
+  resolved "https://registry.yarnpkg.com/@chakra-ui/radio/-/radio-2.0.22.tgz#fad0ce7c9ba4051991ed517cac4cfe526d6d47d9"
+  integrity sha512-GsQ5WAnLwivWl6gPk8P1x+tCcpVakCt5R5T0HumF7DGPXKdJbjS+RaFySrbETmyTJsKY4QrfXn+g8CWVrMjPjw==
   dependencies:
-    "@chakra-ui/form-control" "2.0.17"
-    "@chakra-ui/react-context" "2.0.7"
+    "@chakra-ui/form-control" "2.0.18"
+    "@chakra-ui/react-context" "2.0.8"
     "@chakra-ui/react-types" "2.0.7"
     "@chakra-ui/react-use-merge-refs" "2.0.7"
     "@chakra-ui/shared-utils" "2.0.5"
-    "@zag-js/focus-visible" "0.2.1"
+    "@zag-js/focus-visible" "0.2.2"
 
 "@chakra-ui/react-children-utils@2.0.6":
   version "2.0.6"
   resolved "https://registry.yarnpkg.com/@chakra-ui/react-children-utils/-/react-children-utils-2.0.6.tgz#6c480c6a60678fcb75cb7d57107c7a79e5179b92"
   integrity sha512-QVR2RC7QsOsbWwEnq9YduhpqSFnZGvjjGREV8ygKi8ADhXh93C8azLECCUVgRJF2Wc+So1fgxmjLcbZfY2VmBA==
 
-"@chakra-ui/react-context@2.0.7":
-  version "2.0.7"
-  resolved "https://registry.yarnpkg.com/@chakra-ui/react-context/-/react-context-2.0.7.tgz#f79a2b072d04d4280ec8799dc03a8a1af521ca2e"
-  integrity sha512-i7EGmSU+h2GB30cwrKB4t1R5BMHyGoJM5L2Zz7b+ZUX4aAqyPcfe97wPiQB6Rgr1ImGXrUeov4CDVrRZ2FPgLQ==
+"@chakra-ui/react-context@2.0.8":
+  version "2.0.8"
+  resolved "https://registry.yarnpkg.com/@chakra-ui/react-context/-/react-context-2.0.8.tgz#5e0ed33ac3995875a21dea0e12b0ee5fc4c2e3cc"
+  integrity sha512-tRTKdn6lCTXM6WPjSokAAKCw2ioih7Eg8cNgaYRSwKBck8nkz9YqxgIIEj3dJD7MGtpl24S/SNI98iRWkRwR/A==
 
 "@chakra-ui/react-env@3.0.0":
   version "3.0.0"
@@ -568,12 +568,12 @@
   resolved "https://registry.yarnpkg.com/@chakra-ui/react-use-safe-layout-effect/-/react-use-safe-layout-effect-2.0.5.tgz#6cf388c37fd2a42b5295a292e149b32f860a00a7"
   integrity sha512-MwAQBz3VxoeFLaesaSEN87reVNVbjcQBDex2WGexAg6hUB6n4gc1OWYH/iXp4tzp4kuggBNhEHkk9BMYXWfhJQ==
 
-"@chakra-ui/react-use-size@2.0.9":
-  version "2.0.9"
-  resolved "https://registry.yarnpkg.com/@chakra-ui/react-use-size/-/react-use-size-2.0.9.tgz#00717867b98a24c3bdcfaa0c3e70732404193486"
-  integrity sha512-Jce7QmO1jlQZq+Y77VKckWzroRnajChzUQ8xhLQZO6VbYvrpg3cu+X2QCz3G+MZzB+1/hnvvAqmZ+uJLd8rEJg==
+"@chakra-ui/react-use-size@2.0.10":
+  version "2.0.10"
+  resolved "https://registry.yarnpkg.com/@chakra-ui/react-use-size/-/react-use-size-2.0.10.tgz#6131950852490c06e5fb3760bf64097c8057391f"
+  integrity sha512-fdIkH14GDnKQrtQfxX8N3gxbXRPXEl67Y3zeD9z4bKKcQUAYIMqs0MsPZY+FMpGQw8QqafM44nXfL038aIrC5w==
   dependencies:
-    "@zag-js/element-size" "0.3.1"
+    "@zag-js/element-size" "0.3.2"
 
 "@chakra-ui/react-use-timeout@2.0.5":
   version "2.0.5"
@@ -594,69 +594,69 @@
   dependencies:
     "@chakra-ui/utils" "2.0.15"
 
-"@chakra-ui/react@^2.5.1":
-  version "2.5.1"
-  resolved "https://registry.yarnpkg.com/@chakra-ui/react/-/react-2.5.1.tgz#05414db2b512bd4402e42eecc6b915d85102c576"
-  integrity sha512-ugkaqfcNMb9L4TkalWiF3rnqfr0TlUUD46JZaDIZiORVisaSwXTZTQrVfG40VghhaJT28rnC5WtiE8kd567ZBQ==
+"@chakra-ui/react@^2.5.5":
+  version "2.5.5"
+  resolved "https://registry.yarnpkg.com/@chakra-ui/react/-/react-2.5.5.tgz#5ae2450ec0d10d63e1314747466f21cf542032ff"
+  integrity sha512-aBVMUtdWv2MrptD/tKSqICPsuJ+I+jvauegffO1qPUDlK3RrXIDeOHkLGWohgXNcjY5bGVWguFEzJm97//0ooQ==
   dependencies:
-    "@chakra-ui/accordion" "2.1.9"
-    "@chakra-ui/alert" "2.0.17"
-    "@chakra-ui/avatar" "2.2.5"
-    "@chakra-ui/breadcrumb" "2.1.4"
-    "@chakra-ui/button" "2.0.16"
+    "@chakra-ui/accordion" "2.1.11"
+    "@chakra-ui/alert" "2.1.0"
+    "@chakra-ui/avatar" "2.2.8"
+    "@chakra-ui/breadcrumb" "2.1.5"
+    "@chakra-ui/button" "2.0.18"
     "@chakra-ui/card" "2.1.6"
-    "@chakra-ui/checkbox" "2.2.10"
+    "@chakra-ui/checkbox" "2.2.14"
     "@chakra-ui/close-button" "2.0.17"
     "@chakra-ui/control-box" "2.0.13"
     "@chakra-ui/counter" "2.0.14"
-    "@chakra-ui/css-reset" "2.0.12"
-    "@chakra-ui/editable" "2.0.19"
+    "@chakra-ui/css-reset" "2.1.1"
+    "@chakra-ui/editable" "2.0.21"
     "@chakra-ui/focus-lock" "2.0.16"
-    "@chakra-ui/form-control" "2.0.17"
+    "@chakra-ui/form-control" "2.0.18"
     "@chakra-ui/hooks" "2.1.6"
     "@chakra-ui/icon" "3.0.16"
     "@chakra-ui/image" "2.0.15"
-    "@chakra-ui/input" "2.0.20"
-    "@chakra-ui/layout" "2.1.16"
+    "@chakra-ui/input" "2.0.21"
+    "@chakra-ui/layout" "2.1.18"
     "@chakra-ui/live-region" "2.0.13"
     "@chakra-ui/media-query" "3.2.12"
-    "@chakra-ui/menu" "2.1.9"
-    "@chakra-ui/modal" "2.2.9"
-    "@chakra-ui/number-input" "2.0.18"
-    "@chakra-ui/pin-input" "2.0.19"
-    "@chakra-ui/popover" "2.1.8"
+    "@chakra-ui/menu" "2.1.12"
+    "@chakra-ui/modal" "2.2.11"
+    "@chakra-ui/number-input" "2.0.19"
+    "@chakra-ui/pin-input" "2.0.20"
+    "@chakra-ui/popover" "2.1.9"
     "@chakra-ui/popper" "3.0.13"
-    "@chakra-ui/portal" "2.0.15"
-    "@chakra-ui/progress" "2.1.5"
-    "@chakra-ui/provider" "2.1.2"
-    "@chakra-ui/radio" "2.0.19"
+    "@chakra-ui/portal" "2.0.16"
+    "@chakra-ui/progress" "2.1.6"
+    "@chakra-ui/provider" "2.2.2"
+    "@chakra-ui/radio" "2.0.22"
     "@chakra-ui/react-env" "3.0.0"
-    "@chakra-ui/select" "2.0.18"
+    "@chakra-ui/select" "2.0.19"
     "@chakra-ui/skeleton" "2.0.24"
-    "@chakra-ui/slider" "2.0.21"
+    "@chakra-ui/slider" "2.0.23"
     "@chakra-ui/spinner" "2.0.13"
-    "@chakra-ui/stat" "2.0.17"
-    "@chakra-ui/styled-system" "2.6.1"
-    "@chakra-ui/switch" "2.0.22"
-    "@chakra-ui/system" "2.5.1"
-    "@chakra-ui/table" "2.0.16"
-    "@chakra-ui/tabs" "2.1.8"
-    "@chakra-ui/tag" "2.0.17"
-    "@chakra-ui/textarea" "2.0.18"
-    "@chakra-ui/theme" "2.2.5"
-    "@chakra-ui/theme-utils" "2.0.11"
-    "@chakra-ui/toast" "6.0.1"
-    "@chakra-ui/tooltip" "2.2.6"
-    "@chakra-ui/transition" "2.0.15"
+    "@chakra-ui/stat" "2.0.18"
+    "@chakra-ui/styled-system" "2.8.0"
+    "@chakra-ui/switch" "2.0.26"
+    "@chakra-ui/system" "2.5.5"
+    "@chakra-ui/table" "2.0.17"
+    "@chakra-ui/tabs" "2.1.9"
+    "@chakra-ui/tag" "3.0.0"
+    "@chakra-ui/textarea" "2.0.19"
+    "@chakra-ui/theme" "3.0.1"
+    "@chakra-ui/theme-utils" "2.0.15"
+    "@chakra-ui/toast" "6.1.1"
+    "@chakra-ui/tooltip" "2.2.7"
+    "@chakra-ui/transition" "2.0.16"
     "@chakra-ui/utils" "2.0.15"
     "@chakra-ui/visually-hidden" "2.0.15"
 
-"@chakra-ui/select@2.0.18":
-  version "2.0.18"
-  resolved "https://registry.yarnpkg.com/@chakra-ui/select/-/select-2.0.18.tgz#4eb6092610067c1b4131353fe39b4466e251395b"
-  integrity sha512-1d2lUT5LM6oOs5x4lzBh4GFDuXX62+lr+sgV7099g951/5UNbb0CS2hSZHsO7yZThLNbr7QTWZvAOAayVcGzdw==
+"@chakra-ui/select@2.0.19":
+  version "2.0.19"
+  resolved "https://registry.yarnpkg.com/@chakra-ui/select/-/select-2.0.19.tgz#957e95a17a890d8c0a851e2f00a8d8dd17932d66"
+  integrity sha512-eAlFh+JhwtJ17OrB6fO6gEAGOMH18ERNrXLqWbYLrs674Le7xuREgtuAYDoxUzvYXYYTTdOJtVbcHGriI3o6rA==
   dependencies:
-    "@chakra-ui/form-control" "2.0.17"
+    "@chakra-ui/form-control" "2.0.18"
     "@chakra-ui/shared-utils" "2.0.5"
 
 "@chakra-ui/shared-utils@2.0.5":
@@ -673,20 +673,20 @@
     "@chakra-ui/react-use-previous" "2.0.5"
     "@chakra-ui/shared-utils" "2.0.5"
 
-"@chakra-ui/slider@2.0.21":
-  version "2.0.21"
-  resolved "https://registry.yarnpkg.com/@chakra-ui/slider/-/slider-2.0.21.tgz#f65b15bf0d5f827699ff9a2d6faee35006e2bfce"
-  integrity sha512-Mm76yJxEqJl21+3waEcKg3tM8Y4elJ7mcViN6Brj35PTfzUJfSJxeBGo1nLPJ+X5jLj7o/L4kfBmUk3lY4QYEQ==
+"@chakra-ui/slider@2.0.23":
+  version "2.0.23"
+  resolved "https://registry.yarnpkg.com/@chakra-ui/slider/-/slider-2.0.23.tgz#9130c7aee8ca876be64d1aeba6b84fe421c8207b"
+  integrity sha512-/eyRUXLla+ZdBUPXpakE3SAS2JS8mIJR6qcUYiPVKSpRAi6tMyYeQijAXn2QC1AUVd2JrG8Pz+1Jy7Po3uA7cA==
   dependencies:
     "@chakra-ui/number-utils" "2.0.7"
-    "@chakra-ui/react-context" "2.0.7"
+    "@chakra-ui/react-context" "2.0.8"
     "@chakra-ui/react-types" "2.0.7"
     "@chakra-ui/react-use-callback-ref" "2.0.7"
     "@chakra-ui/react-use-controllable-state" "2.0.8"
     "@chakra-ui/react-use-latest-ref" "2.0.5"
     "@chakra-ui/react-use-merge-refs" "2.0.7"
     "@chakra-ui/react-use-pan-event" "2.0.9"
-    "@chakra-ui/react-use-size" "2.0.9"
+    "@chakra-ui/react-use-size" "2.0.10"
     "@chakra-ui/react-use-update-effect" "2.0.7"
 
 "@chakra-ui/spinner@2.0.13":
@@ -696,82 +696,82 @@
   dependencies:
     "@chakra-ui/shared-utils" "2.0.5"
 
-"@chakra-ui/stat@2.0.17":
-  version "2.0.17"
-  resolved "https://registry.yarnpkg.com/@chakra-ui/stat/-/stat-2.0.17.tgz#2cd712cc7e0d58d9cbd542deea911f1b0925074f"
-  integrity sha512-PhD+5oVLWjQmGLfeZSmexp3AtLcaggWBwoMZ4z8QMZIQzf/fJJWMk0bMqxlpTv8ORDkfY/4ImuFB/RJHvcqlcA==
+"@chakra-ui/stat@2.0.18":
+  version "2.0.18"
+  resolved "https://registry.yarnpkg.com/@chakra-ui/stat/-/stat-2.0.18.tgz#9e5d21d162b7cf2cf92065c19291ead2d4660772"
+  integrity sha512-wKyfBqhVlIs9bkSerUc6F9KJMw0yTIEKArW7dejWwzToCLPr47u+CtYO6jlJHV6lRvkhi4K4Qc6pyvtJxZ3VpA==
   dependencies:
     "@chakra-ui/icon" "3.0.16"
-    "@chakra-ui/react-context" "2.0.7"
+    "@chakra-ui/react-context" "2.0.8"
     "@chakra-ui/shared-utils" "2.0.5"
 
-"@chakra-ui/styled-system@2.6.1":
-  version "2.6.1"
-  resolved "https://registry.yarnpkg.com/@chakra-ui/styled-system/-/styled-system-2.6.1.tgz#302d496d34c0b7b30c646a7e3c9b113a2f4588da"
-  integrity sha512-jy/1dVi1LxjoRCm+Eo5mqBgvPy5SCWMlIcz6GbIZBDpkGeKZwtqrZLjekxxLBCy8ORY+kJlUB0FT6AzVR/1tjw==
+"@chakra-ui/styled-system@2.8.0":
+  version "2.8.0"
+  resolved "https://registry.yarnpkg.com/@chakra-ui/styled-system/-/styled-system-2.8.0.tgz#c02aa7b4a15bd826c19d055cd226bd44f7470f26"
+  integrity sha512-bmRv/8ACJGGKGx84U1npiUddwdNifJ+/ETklGwooS5APM0ymwUtBYZpFxjYNJrqvVYpg3mVY6HhMyBVptLS7iA==
   dependencies:
     "@chakra-ui/shared-utils" "2.0.5"
     csstype "^3.0.11"
     lodash.mergewith "4.6.2"
 
-"@chakra-ui/switch@2.0.22":
-  version "2.0.22"
-  resolved "https://registry.yarnpkg.com/@chakra-ui/switch/-/switch-2.0.22.tgz#7b35e2b10ea4cf91fb49f5175b4335c61dcd25b3"
-  integrity sha512-+/Yy6y7VFD91uSPruF8ZvePi3tl5D8UNVATtWEQ+QBI92DLSM+PtgJ2F0Y9GMZ9NzMxpZ80DqwY7/kqcPCfLvw==
+"@chakra-ui/switch@2.0.26":
+  version "2.0.26"
+  resolved "https://registry.yarnpkg.com/@chakra-ui/switch/-/switch-2.0.26.tgz#b93eeafd788e47c21222524adceffe9ef62602d6"
+  integrity sha512-x62lF6VazSZJQuVxosChVR6+0lIJe8Pxgkl/C9vxjhp2yVYb3mew5tcX/sDOu0dYZy8ro/9hMfGkdN4r9xEU8A==
   dependencies:
-    "@chakra-ui/checkbox" "2.2.10"
+    "@chakra-ui/checkbox" "2.2.14"
     "@chakra-ui/shared-utils" "2.0.5"
 
-"@chakra-ui/system@2.5.1":
-  version "2.5.1"
-  resolved "https://registry.yarnpkg.com/@chakra-ui/system/-/system-2.5.1.tgz#bc03a11ae31e795966c7618280548d5cd866f47e"
-  integrity sha512-4+86OrcSoq7lGkm5fh+sJ3IWXSTzjz+HOllRbCW2Rtnmcg7ritiXVNV2VygEg2DrCcx5+tNqRHDM764zW+AEug==
+"@chakra-ui/system@2.5.5":
+  version "2.5.5"
+  resolved "https://registry.yarnpkg.com/@chakra-ui/system/-/system-2.5.5.tgz#b8b070d07ca9b0190363100396eea02cca754cec"
+  integrity sha512-52BIp/Zyvefgxn5RTByfkTeG4J+y81LWEjWm8jCaRFsLVm8IFgqIrngtcq4I7gD5n/UKbneHlb4eLHo4uc5yDQ==
   dependencies:
     "@chakra-ui/color-mode" "2.1.12"
     "@chakra-ui/object-utils" "2.0.8"
     "@chakra-ui/react-utils" "2.0.12"
-    "@chakra-ui/styled-system" "2.6.1"
-    "@chakra-ui/theme-utils" "2.0.11"
+    "@chakra-ui/styled-system" "2.8.0"
+    "@chakra-ui/theme-utils" "2.0.15"
     "@chakra-ui/utils" "2.0.15"
-    react-fast-compare "3.2.0"
+    react-fast-compare "3.2.1"
 
-"@chakra-ui/table@2.0.16":
-  version "2.0.16"
-  resolved "https://registry.yarnpkg.com/@chakra-ui/table/-/table-2.0.16.tgz#e69736cba5cfb218c5e40592ad9280c6e32f6fe7"
-  integrity sha512-vWDXZ6Ad3Aj66curp1tZBHvCfQHX2FJ4ijLiqGgQszWFIchfhJ5vMgEBJaFMZ+BN1draAjuRTZqaQefOApzvRg==
+"@chakra-ui/table@2.0.17":
+  version "2.0.17"
+  resolved "https://registry.yarnpkg.com/@chakra-ui/table/-/table-2.0.17.tgz#ad394dc6dcbe5a8a9e6d899997ecca3471603977"
+  integrity sha512-OScheTEp1LOYvTki2NFwnAYvac8siAhW9BI5RKm5f5ORL2gVJo4I72RUqE0aKe1oboxgm7CYt5afT5PS5cG61A==
   dependencies:
-    "@chakra-ui/react-context" "2.0.7"
+    "@chakra-ui/react-context" "2.0.8"
     "@chakra-ui/shared-utils" "2.0.5"
 
-"@chakra-ui/tabs@2.1.8":
-  version "2.1.8"
-  resolved "https://registry.yarnpkg.com/@chakra-ui/tabs/-/tabs-2.1.8.tgz#e83071380f9a3633810308d45de51be7a74f5eb9"
-  integrity sha512-B7LeFN04Ny2jsSy5TFOQxnbZ6ITxGxLxsB2PE0vvQjMSblBrUryOxdjw80HZhfiw6od0ikK9CeKQOIt9QCguSw==
+"@chakra-ui/tabs@2.1.9":
+  version "2.1.9"
+  resolved "https://registry.yarnpkg.com/@chakra-ui/tabs/-/tabs-2.1.9.tgz#2e5214cb453c6cc0c240e82bd88af1042fc6fe0e"
+  integrity sha512-Yf8e0kRvaGM6jfkJum0aInQ0U3ZlCafmrYYni2lqjcTtThqu+Yosmo3iYlnullXxCw5MVznfrkb9ySvgQowuYg==
   dependencies:
     "@chakra-ui/clickable" "2.0.14"
-    "@chakra-ui/descendant" "3.0.13"
+    "@chakra-ui/descendant" "3.0.14"
     "@chakra-ui/lazy-utils" "2.0.5"
     "@chakra-ui/react-children-utils" "2.0.6"
-    "@chakra-ui/react-context" "2.0.7"
+    "@chakra-ui/react-context" "2.0.8"
     "@chakra-ui/react-use-controllable-state" "2.0.8"
     "@chakra-ui/react-use-merge-refs" "2.0.7"
     "@chakra-ui/react-use-safe-layout-effect" "2.0.5"
     "@chakra-ui/shared-utils" "2.0.5"
 
-"@chakra-ui/tag@2.0.17":
-  version "2.0.17"
-  resolved "https://registry.yarnpkg.com/@chakra-ui/tag/-/tag-2.0.17.tgz#97adb86db190ddb3526060b78c590392e0ac8b4c"
-  integrity sha512-A47zE9Ft9qxOJ+5r1cUseKRCoEdqCRzFm0pOtZgRcckqavglk75Xjgz8HbBpUO2zqqd49MlqdOwR8o87fXS1vg==
+"@chakra-ui/tag@3.0.0":
+  version "3.0.0"
+  resolved "https://registry.yarnpkg.com/@chakra-ui/tag/-/tag-3.0.0.tgz#d86cdab59bb3ff7fc628c2dbe7a5ff1b36bd3e96"
+  integrity sha512-YWdMmw/1OWRwNkG9pX+wVtZio+B89odaPj6XeMn5nfNN8+jyhIEpouWv34+CO9G0m1lupJTxPSfgLAd7cqXZMA==
   dependencies:
     "@chakra-ui/icon" "3.0.16"
-    "@chakra-ui/react-context" "2.0.7"
+    "@chakra-ui/react-context" "2.0.8"
 
-"@chakra-ui/textarea@2.0.18":
-  version "2.0.18"
-  resolved "https://registry.yarnpkg.com/@chakra-ui/textarea/-/textarea-2.0.18.tgz#da6d629b465f65bbc7b48039c2e48a4ae1d853b4"
-  integrity sha512-aGHHb29vVifO0OtcK/k8cMykzjOKo/coDTU0NJqz7OOLAWIMNV2eGenvmO1n9tTZbmbqHiX+Sa1nPRX+pd14lg==
+"@chakra-ui/textarea@2.0.19":
+  version "2.0.19"
+  resolved "https://registry.yarnpkg.com/@chakra-ui/textarea/-/textarea-2.0.19.tgz#470b459f9cb3255d2abbe07d46b0a5b60a6a32c5"
+  integrity sha512-adJk+qVGsFeJDvfn56CcJKKse8k7oMGlODrmpnpTdF+xvlsiTM+1GfaJvgNSpHHuQFdz/A0z1uJtfGefk0G2ZA==
   dependencies:
-    "@chakra-ui/form-control" "2.0.17"
+    "@chakra-ui/form-control" "2.0.18"
     "@chakra-ui/shared-utils" "2.0.5"
 
 "@chakra-ui/theme-tools@2.0.17":
@@ -783,57 +783,57 @@
     "@chakra-ui/shared-utils" "2.0.5"
     color2k "^2.0.0"
 
-"@chakra-ui/theme-utils@2.0.11":
-  version "2.0.11"
-  resolved "https://registry.yarnpkg.com/@chakra-ui/theme-utils/-/theme-utils-2.0.11.tgz#c01b1d14fdd63326d1ad11fd8f0872921ea43872"
-  integrity sha512-lBAay6Sq3/fl7exd3mFxWAbzgdQowytor0fnlHrpNStn1HgFjXukwsf6356XQOie2Vd8qaMM7qZtMh4AiC0dcg==
+"@chakra-ui/theme-utils@2.0.15":
+  version "2.0.15"
+  resolved "https://registry.yarnpkg.com/@chakra-ui/theme-utils/-/theme-utils-2.0.15.tgz#968a5e8c47bb403323fe67049c7b751a6e47f069"
+  integrity sha512-UuxtEgE7gwMTGDXtUpTOI7F5X0iHB9ekEOG5PWPn2wWBL7rlk2JtPI7UP5Um5Yg6vvBfXYGK1ySahxqsgf+87g==
   dependencies:
     "@chakra-ui/shared-utils" "2.0.5"
-    "@chakra-ui/styled-system" "2.6.1"
-    "@chakra-ui/theme" "2.2.5"
+    "@chakra-ui/styled-system" "2.8.0"
+    "@chakra-ui/theme" "3.0.1"
     lodash.mergewith "4.6.2"
 
-"@chakra-ui/theme@2.2.5":
-  version "2.2.5"
-  resolved "https://registry.yarnpkg.com/@chakra-ui/theme/-/theme-2.2.5.tgz#18ed1755ff27c1ff1f1a77083ffc546c361c926e"
-  integrity sha512-hYASZMwu0NqEv6PPydu+F3I+kMNd44yR4TwjR/lXBz/LEh64L6UPY6kQjebCfgdVtsGdl3HKg+eLlfa7SvfRgw==
+"@chakra-ui/theme@3.0.1":
+  version "3.0.1"
+  resolved "https://registry.yarnpkg.com/@chakra-ui/theme/-/theme-3.0.1.tgz#151fc5d1e23d0fd0cd29d28acf8f6017269c13fc"
+  integrity sha512-92kDm/Ux/51uJqhRKevQo/O/rdwucDYcpHg2QuwzdAxISCeYvgtl2TtgOOl5EnqEP0j3IEAvZHZUlv8TTbawaw==
   dependencies:
     "@chakra-ui/anatomy" "2.1.2"
     "@chakra-ui/shared-utils" "2.0.5"
     "@chakra-ui/theme-tools" "2.0.17"
 
-"@chakra-ui/toast@6.0.1":
-  version "6.0.1"
-  resolved "https://registry.yarnpkg.com/@chakra-ui/toast/-/toast-6.0.1.tgz#726b67a57cdd592320bb3f450c66d007a2a1d902"
-  integrity sha512-ej2kJXvu/d2h6qnXU5D8XTyw0qpsfmbiU7hUffo/sPxkz89AUOQ08RUuUmB1ssW/FZcQvNMJ5WgzCTKHGBxtxw==
+"@chakra-ui/toast@6.1.1":
+  version "6.1.1"
+  resolved "https://registry.yarnpkg.com/@chakra-ui/toast/-/toast-6.1.1.tgz#7ca78f38069bc87fa75b64de76c8fc2758bdf419"
+  integrity sha512-JtjIKkPVjEu8okGGCipCxNVgK/15h5AicTATZ6RbG2MsHmr4GfKG3fUCvpbuZseArqmLqGLQZQJjVE9vJzaSkQ==
   dependencies:
-    "@chakra-ui/alert" "2.0.17"
+    "@chakra-ui/alert" "2.1.0"
     "@chakra-ui/close-button" "2.0.17"
-    "@chakra-ui/portal" "2.0.15"
-    "@chakra-ui/react-context" "2.0.7"
+    "@chakra-ui/portal" "2.0.16"
+    "@chakra-ui/react-context" "2.0.8"
     "@chakra-ui/react-use-timeout" "2.0.5"
     "@chakra-ui/react-use-update-effect" "2.0.7"
     "@chakra-ui/shared-utils" "2.0.5"
-    "@chakra-ui/styled-system" "2.6.1"
-    "@chakra-ui/theme" "2.2.5"
+    "@chakra-ui/styled-system" "2.8.0"
+    "@chakra-ui/theme" "3.0.1"
 
-"@chakra-ui/tooltip@2.2.6":
-  version "2.2.6"
-  resolved "https://registry.yarnpkg.com/@chakra-ui/tooltip/-/tooltip-2.2.6.tgz#a38f9ff2dd8a574c8cf49526c3846533455f8ddd"
-  integrity sha512-4cbneidZ5+HCWge3OZzewRQieIvhDjSsl+scrl4Scx7E0z3OmqlTIESU5nGIZDBLYqKn/UirEZhqaQ33FOS2fw==
+"@chakra-ui/tooltip@2.2.7":
+  version "2.2.7"
+  resolved "https://registry.yarnpkg.com/@chakra-ui/tooltip/-/tooltip-2.2.7.tgz#7c305efb057a5fe4694b1b8d82395aec776d8f57"
+  integrity sha512-ImUJ6NnVqARaYqpgtO+kzucDRmxo8AF3jMjARw0bx2LxUkKwgRCOEaaRK5p5dHc0Kr6t5/XqjDeUNa19/sLauA==
   dependencies:
     "@chakra-ui/popper" "3.0.13"
-    "@chakra-ui/portal" "2.0.15"
+    "@chakra-ui/portal" "2.0.16"
     "@chakra-ui/react-types" "2.0.7"
     "@chakra-ui/react-use-disclosure" "2.0.8"
     "@chakra-ui/react-use-event-listener" "2.0.7"
     "@chakra-ui/react-use-merge-refs" "2.0.7"
     "@chakra-ui/shared-utils" "2.0.5"
 
-"@chakra-ui/transition@2.0.15":
-  version "2.0.15"
-  resolved "https://registry.yarnpkg.com/@chakra-ui/transition/-/transition-2.0.15.tgz#c640df2ea82f5ad58c55a6e1a7c338f377cb96d8"
-  integrity sha512-o9LBK/llQfUDHF/Ty3cQ6nShpekKTqHUoJlUOzNKhoTsNpoRerr9v0jwojrX1YI02KtVjfhFU6PiqXlDfREoNw==
+"@chakra-ui/transition@2.0.16":
+  version "2.0.16"
+  resolved "https://registry.yarnpkg.com/@chakra-ui/transition/-/transition-2.0.16.tgz#498c91e6835bb5d950fd1d1402f483b85f7dcd87"
+  integrity sha512-E+RkwlPc3H7P1crEXmXwDXMB2lqY2LLia2P5siQ4IEnRWIgZXlIw+8Em+NtHNgusel2N+9yuB0wT9SeZZeZ3CQ==
   dependencies:
     "@chakra-ui/shared-utils" "2.0.5"
 
@@ -1081,6 +1081,18 @@
   resolved "https://registry.yarnpkg.com/@esbuild/win32-x64/-/win32-x64-0.16.17.tgz#c5a1a4bfe1b57f0c3e61b29883525c6da3e5c091"
   integrity sha512-y+EHuSchhL7FjHgvQL/0fnnFmO4T1bhvWANX6gcnqTjtnKWbTvUMCpGnv2+t+31d7RzyEAYAd4u2fnIhHL6N/Q==
 
+"@eslint-community/eslint-utils@^4.2.0":
+  version "4.4.0"
+  resolved "https://registry.yarnpkg.com/@eslint-community/eslint-utils/-/eslint-utils-4.4.0.tgz#a23514e8fb9af1269d5f7788aa556798d61c6b59"
+  integrity sha512-1/sA4dwrzBAyeUoQ6oxahHKmrZvsnLCg4RfxW3ZFGGmQkSNQPFNLV9CUEFQP1x9EYXHTo5p6xdhZM1Ne9p/AfA==
+  dependencies:
+    eslint-visitor-keys "^3.3.0"
+
+"@eslint-community/regexpp@^4.4.0":
+  version "4.5.0"
+  resolved "https://registry.yarnpkg.com/@eslint-community/regexpp/-/regexpp-4.5.0.tgz#f6f729b02feee2c749f57e334b7a1b5f40a81724"
+  integrity sha512-vITaYzIcNmjn5tF5uxcZ/ft7/RXGrMUIS9HalWckEOF6ESiwXKoMzAQf2UW0aVd6rnOeExTJVd5hmWXucBKGXQ==
+
 "@eslint/eslintrc@^1.4.1":
   version "1.4.1"
   resolved "https://registry.yarnpkg.com/@eslint/eslintrc/-/eslintrc-1.4.1.tgz#af58772019a2d271b7e2d4c23ff4ddcba3ccfb3e"
@@ -1768,47 +1780,47 @@
   resolved "https://registry.yarnpkg.com/@types/uuid/-/uuid-9.0.0.tgz#53ef263e5239728b56096b0a869595135b7952d2"
   integrity sha512-kr90f+ERiQtKWMz5rP32ltJ/BtULDI5RVO0uavn1HQUOwjx0R1h0rnDYNL0CepF1zL5bSY6FISAfd9tOdDhU5Q==
 
-"@typescript-eslint/eslint-plugin@^5.52.0":
-  version "5.52.0"
-  resolved "https://registry.yarnpkg.com/@typescript-eslint/eslint-plugin/-/eslint-plugin-5.52.0.tgz#5fb0d43574c2411f16ea80f5fc335b8eaa7b28a8"
-  integrity sha512-lHazYdvYVsBokwCdKOppvYJKaJ4S41CgKBcPvyd0xjZNbvQdhn/pnJlGtQksQ/NhInzdaeaSarlBjDXHuclEbg==
+"@typescript-eslint/eslint-plugin@^5.57.0":
+  version "5.57.0"
+  resolved "https://registry.yarnpkg.com/@typescript-eslint/eslint-plugin/-/eslint-plugin-5.57.0.tgz#52c8a7a4512f10e7249ca1e2e61f81c62c34365c"
+  integrity sha512-itag0qpN6q2UMM6Xgk6xoHa0D0/P+M17THnr4SVgqn9Rgam5k/He33MA7/D7QoJcdMxHFyX7U9imaBonAX/6qA==
   dependencies:
-    "@typescript-eslint/scope-manager" "5.52.0"
-    "@typescript-eslint/type-utils" "5.52.0"
-    "@typescript-eslint/utils" "5.52.0"
+    "@eslint-community/regexpp" "^4.4.0"
+    "@typescript-eslint/scope-manager" "5.57.0"
+    "@typescript-eslint/type-utils" "5.57.0"
+    "@typescript-eslint/utils" "5.57.0"
     debug "^4.3.4"
     grapheme-splitter "^1.0.4"
     ignore "^5.2.0"
     natural-compare-lite "^1.4.0"
-    regexpp "^3.2.0"
     semver "^7.3.7"
     tsutils "^3.21.0"
 
-"@typescript-eslint/parser@^5.52.0":
-  version "5.52.0"
-  resolved "https://registry.yarnpkg.com/@typescript-eslint/parser/-/parser-5.52.0.tgz#73c136df6c0133f1d7870de7131ccf356f5be5a4"
-  integrity sha512-e2KiLQOZRo4Y0D/b+3y08i3jsekoSkOYStROYmPUnGMEoA0h+k2qOH5H6tcjIc68WDvGwH+PaOrP1XRzLJ6QlA==
+"@typescript-eslint/parser@^5.57.0":
+  version "5.57.0"
+  resolved "https://registry.yarnpkg.com/@typescript-eslint/parser/-/parser-5.57.0.tgz#f675bf2cd1a838949fd0de5683834417b757e4fa"
+  integrity sha512-orrduvpWYkgLCyAdNtR1QIWovcNZlEm6yL8nwH/eTxWLd8gsP+25pdLHYzL2QdkqrieaDwLpytHqycncv0woUQ==
   dependencies:
-    "@typescript-eslint/scope-manager" "5.52.0"
-    "@typescript-eslint/types" "5.52.0"
-    "@typescript-eslint/typescript-estree" "5.52.0"
+    "@typescript-eslint/scope-manager" "5.57.0"
+    "@typescript-eslint/types" "5.57.0"
+    "@typescript-eslint/typescript-estree" "5.57.0"
     debug "^4.3.4"
 
-"@typescript-eslint/scope-manager@5.52.0":
-  version "5.52.0"
-  resolved "https://registry.yarnpkg.com/@typescript-eslint/scope-manager/-/scope-manager-5.52.0.tgz#a993d89a0556ea16811db48eabd7c5b72dcb83d1"
-  integrity sha512-AR7sxxfBKiNV0FWBSARxM8DmNxrwgnYMPwmpkC1Pl1n+eT8/I2NAUPuwDy/FmDcC6F8pBfmOcaxcxRHspgOBMw==
+"@typescript-eslint/scope-manager@5.57.0":
+  version "5.57.0"
+  resolved "https://registry.yarnpkg.com/@typescript-eslint/scope-manager/-/scope-manager-5.57.0.tgz#79ccd3fa7bde0758059172d44239e871e087ea36"
+  integrity sha512-NANBNOQvllPlizl9LatX8+MHi7bx7WGIWYjPHDmQe5Si/0YEYfxSljJpoTyTWFTgRy3X8gLYSE4xQ2U+aCozSw==
   dependencies:
-    "@typescript-eslint/types" "5.52.0"
-    "@typescript-eslint/visitor-keys" "5.52.0"
+    "@typescript-eslint/types" "5.57.0"
+    "@typescript-eslint/visitor-keys" "5.57.0"
 
-"@typescript-eslint/type-utils@5.52.0":
-  version "5.52.0"
-  resolved "https://registry.yarnpkg.com/@typescript-eslint/type-utils/-/type-utils-5.52.0.tgz#9fd28cd02e6f21f5109e35496df41893f33167aa"
-  integrity sha512-tEKuUHfDOv852QGlpPtB3lHOoig5pyFQN/cUiZtpw99D93nEBjexRLre5sQZlkMoHry/lZr8qDAt2oAHLKA6Jw==
+"@typescript-eslint/type-utils@5.57.0":
+  version "5.57.0"
+  resolved "https://registry.yarnpkg.com/@typescript-eslint/type-utils/-/type-utils-5.57.0.tgz#98e7531c4e927855d45bd362de922a619b4319f2"
+  integrity sha512-kxXoq9zOTbvqzLbdNKy1yFrxLC6GDJFE2Yuo3KqSwTmDOFjUGeWSakgoXT864WcK5/NAJkkONCiKb1ddsqhLXQ==
   dependencies:
-    "@typescript-eslint/typescript-estree" "5.52.0"
-    "@typescript-eslint/utils" "5.52.0"
+    "@typescript-eslint/typescript-estree" "5.57.0"
+    "@typescript-eslint/utils" "5.57.0"
     debug "^4.3.4"
     tsutils "^3.21.0"
 
@@ -1822,13 +1834,18 @@
   resolved "https://registry.yarnpkg.com/@typescript-eslint/types/-/types-5.52.0.tgz#19e9abc6afb5bd37a1a9bea877a1a836c0b3241b"
   integrity sha512-oV7XU4CHYfBhk78fS7tkum+/Dpgsfi91IIDy7fjCyq2k6KB63M6gMC0YIvy+iABzmXThCRI6xpCEyVObBdWSDQ==
 
-"@typescript-eslint/typescript-estree@5.52.0", "@typescript-eslint/typescript-estree@^5.13.0":
-  version "5.52.0"
-  resolved "https://registry.yarnpkg.com/@typescript-eslint/typescript-estree/-/typescript-estree-5.52.0.tgz#6408cb3c2ccc01c03c278cb201cf07e73347dfca"
-  integrity sha512-WeWnjanyEwt6+fVrSR0MYgEpUAuROxuAH516WPjUblIrClzYJj0kBbjdnbQXLpgAN8qbEuGywiQsXUVDiAoEuQ==
+"@typescript-eslint/types@5.57.0":
+  version "5.57.0"
+  resolved "https://registry.yarnpkg.com/@typescript-eslint/types/-/types-5.57.0.tgz#727bfa2b64c73a4376264379cf1f447998eaa132"
+  integrity sha512-mxsod+aZRSyLT+jiqHw1KK6xrANm19/+VFALVFP5qa/aiJnlP38qpyaTd0fEKhWvQk6YeNZ5LGwI1pDpBRBhtQ==
+
+"@typescript-eslint/typescript-estree@5.57.0":
+  version "5.57.0"
+  resolved "https://registry.yarnpkg.com/@typescript-eslint/typescript-estree/-/typescript-estree-5.57.0.tgz#ebcd0ee3e1d6230e888d88cddf654252d41e2e40"
+  integrity sha512-LTzQ23TV82KpO8HPnWuxM2V7ieXW8O142I7hQTxWIHDcCEIjtkat6H96PFkYBQqGFLW/G/eVVOB9Z8rcvdY/Vw==
   dependencies:
-    "@typescript-eslint/types" "5.52.0"
-    "@typescript-eslint/visitor-keys" "5.52.0"
+    "@typescript-eslint/types" "5.57.0"
+    "@typescript-eslint/visitor-keys" "5.57.0"
     debug "^4.3.4"
     globby "^11.1.0"
     is-glob "^4.0.3"
@@ -1848,18 +1865,31 @@
     semver "^7.3.5"
     tsutils "^3.21.0"
 
-"@typescript-eslint/utils@5.52.0":
+"@typescript-eslint/typescript-estree@^5.13.0":
   version "5.52.0"
-  resolved "https://registry.yarnpkg.com/@typescript-eslint/utils/-/utils-5.52.0.tgz#b260bb5a8f6b00a0ed51db66bdba4ed5e4845a72"
-  integrity sha512-As3lChhrbwWQLNk2HC8Ree96hldKIqk98EYvypd3It8Q1f8d5zWyIoaZEp2va5667M4ZyE7X8UUR+azXrFl+NA==
+  resolved "https://registry.yarnpkg.com/@typescript-eslint/typescript-estree/-/typescript-estree-5.52.0.tgz#6408cb3c2ccc01c03c278cb201cf07e73347dfca"
+  integrity sha512-WeWnjanyEwt6+fVrSR0MYgEpUAuROxuAH516WPjUblIrClzYJj0kBbjdnbQXLpgAN8qbEuGywiQsXUVDiAoEuQ==
   dependencies:
+    "@typescript-eslint/types" "5.52.0"
+    "@typescript-eslint/visitor-keys" "5.52.0"
+    debug "^4.3.4"
+    globby "^11.1.0"
+    is-glob "^4.0.3"
+    semver "^7.3.7"
+    tsutils "^3.21.0"
+
+"@typescript-eslint/utils@5.57.0":
+  version "5.57.0"
+  resolved "https://registry.yarnpkg.com/@typescript-eslint/utils/-/utils-5.57.0.tgz#eab8f6563a2ac31f60f3e7024b91bf75f43ecef6"
+  integrity sha512-ps/4WohXV7C+LTSgAL5CApxvxbMkl9B9AUZRtnEFonpIxZDIT7wC1xfvuJONMidrkB9scs4zhtRyIwHh4+18kw==
+  dependencies:
+    "@eslint-community/eslint-utils" "^4.2.0"
     "@types/json-schema" "^7.0.9"
     "@types/semver" "^7.3.12"
-    "@typescript-eslint/scope-manager" "5.52.0"
-    "@typescript-eslint/types" "5.52.0"
-    "@typescript-eslint/typescript-estree" "5.52.0"
+    "@typescript-eslint/scope-manager" "5.57.0"
+    "@typescript-eslint/types" "5.57.0"
+    "@typescript-eslint/typescript-estree" "5.57.0"
     eslint-scope "^5.1.1"
-    eslint-utils "^3.0.0"
     semver "^7.3.7"
 
 "@typescript-eslint/visitor-keys@4.33.0":
@@ -1878,6 +1908,14 @@
     "@typescript-eslint/types" "5.52.0"
     eslint-visitor-keys "^3.3.0"
 
+"@typescript-eslint/visitor-keys@5.57.0":
+  version "5.57.0"
+  resolved "https://registry.yarnpkg.com/@typescript-eslint/visitor-keys/-/visitor-keys-5.57.0.tgz#e2b2f4174aff1d15eef887ce3d019ecc2d7a8ac1"
+  integrity sha512-ery2g3k0hv5BLiKpPuwYt9KBkAp2ugT6VvyShXdLOkax895EC55sP0Tx5L0fZaQueiK3fBLvHVvEl3jFS5ia+g==
+  dependencies:
+    "@typescript-eslint/types" "5.57.0"
+    eslint-visitor-keys "^3.3.0"
+
 "@vitejs/plugin-react-swc@^3.2.0":
   version "3.2.0"
   resolved "https://registry.yarnpkg.com/@vitejs/plugin-react-swc/-/plugin-react-swc-3.2.0.tgz#7c4f6e116a296c27f680d05750f9dbf798cf7709"
@@ -1890,15 +1928,15 @@
   resolved "https://registry.yarnpkg.com/@yarnpkg/lockfile/-/lockfile-1.1.0.tgz#e77a97fbd345b76d83245edcd17d393b1b41fb31"
   integrity sha512-GpSwvyXOcOOlV70vbnzjj4fW5xW/FdUF6nQEt1ENy7m4ZCczi1+/buVUPAqmGfqznsORNFzUMjctTIp8a9tuCQ==
 
-"@zag-js/element-size@0.3.1":
-  version "0.3.1"
-  resolved "https://registry.yarnpkg.com/@zag-js/element-size/-/element-size-0.3.1.tgz#f9f6ae98355e2250d18d0f6e2f1134a0ae4c6a2f"
-  integrity sha512-jR5j4G//bRzcxwAACWi9EfITnwjNmn10LxF4NmALrdZU7/PNWP3uUCdhCxd/0SCyeiJXUl0yvD57rWAbKPs1nw==
+"@zag-js/element-size@0.3.2":
+  version "0.3.2"
+  resolved "https://registry.yarnpkg.com/@zag-js/element-size/-/element-size-0.3.2.tgz#ebb76af2a024230482406db41344598d1a9f54f4"
+  integrity sha512-bVvvigUGvAuj7PCkE5AbzvTJDTw5f3bg9nQdv+ErhVN8SfPPppLJEmmWdxqsRzrHXgx8ypJt/+Ty0kjtISVDsQ==
 
-"@zag-js/focus-visible@0.2.1":
-  version "0.2.1"
-  resolved "https://registry.yarnpkg.com/@zag-js/focus-visible/-/focus-visible-0.2.1.tgz#bf4f1009f4fd35a9728dfaa9214d8cb318fe8b1e"
-  integrity sha512-19uTjoZGP4/Ax7kSNhhay9JA83BirKzpqLkeEAilrpdI1hE5xuq6q+tzJOsrMOOqJrm7LkmZp5lbsTQzvK2pYg==
+"@zag-js/focus-visible@0.2.2":
+  version "0.2.2"
+  resolved "https://registry.yarnpkg.com/@zag-js/focus-visible/-/focus-visible-0.2.2.tgz#56233480ca1275d3218fb2e10696a33d1a6b9e64"
+  integrity sha512-0j2gZq8HiZ51z4zNnSkF1iSkqlwRDvdH+son3wHdoz+7IUdMN/5Exd4TxMJ+gq2Of1DiXReYLL9qqh2PdQ4wgA==
 
 accepts@~1.3.4:
   version "1.3.8"
@@ -4510,10 +4548,10 @@ react-dropzone@^14.2.3:
     file-selector "^0.6.0"
     prop-types "^15.8.1"
 
-react-fast-compare@3.2.0:
-  version "3.2.0"
-  resolved "https://registry.yarnpkg.com/react-fast-compare/-/react-fast-compare-3.2.0.tgz#641a9da81b6a6320f270e89724fb45a0b39e43bb"
-  integrity sha512-rtGImPZ0YyLrscKI9xTpV8psd6I8VAtjKCzQDlzyDvqJA8XOW78TXYQwNRNd8g8JZnDu8q9Fu/1v4HPAVwVdHA==
+react-fast-compare@3.2.1:
+  version "3.2.1"
+  resolved "https://registry.yarnpkg.com/react-fast-compare/-/react-fast-compare-3.2.1.tgz#53933d9e14f364281d6cba24bfed7a4afb808b5f"
+  integrity sha512-xTYf9zFim2pEif/Fw16dBiXpe0hoy5PxcD8+OwBnTtNLfIm3g6WxhKNurY+6OmdH1u6Ta/W/Vl6vjbYP1MFnDg==
 
 react-fast-compare@^2.0.1:
   version "2.0.4"
@@ -5311,6 +5349,11 @@ typescript@^4.0.0, typescript@^4.5.5:
   resolved "https://registry.yarnpkg.com/typescript/-/typescript-4.9.5.tgz#095979f9bcc0d09da324d58d03ce8f8374cbe65a"
   integrity sha512-1FXk9E2Hm+QzZQ7z+McJiHL4NW1F2EzMu9Nq9i3zAaGqibafqYwCVU6WyWAuyQRRzOlxou8xZSyXLEN8oKj24g==
 
+typescript@^5.0.3:
+  version "5.0.3"
+  resolved "https://registry.yarnpkg.com/typescript/-/typescript-5.0.3.tgz#fe976f0c826a88d0a382007681cbb2da44afdedf"
+  integrity sha512-xv8mOEDnigb/tN9PSMTwSEqAnUvkoXMQlicOb0IUVDBSQCgBSaAAROUZYy2IcUy5qU6XajK5jjjO7TMWqBTKZA==
+
 unbox-primitive@^1.0.2:
   version "1.0.2"
   resolved "https://registry.yarnpkg.com/unbox-primitive/-/unbox-primitive-1.0.2.tgz#29032021057d5e6cdbd08c5129c226dff8ed6f9e"

ldm/generate.py

@@ -22,6 +22,7 @@ import transformers
 from diffusers.pipeline_utils import DiffusionPipeline
 from diffusers.utils.import_utils import is_xformers_available
 from omegaconf import OmegaConf
+from pathlib import Path
 from PIL import Image, ImageOps
 from pytorch_lightning import logging, seed_everything
 
@@ -528,6 +529,11 @@ class Generate:
                 log_tokens=self.log_tokenization,
             )
 
+
+            # untested code, so commented out for now
+            # if self.model.peft_manager:
+            #     self.model = self.model.peft_manager.load(self.model, self.model.unet.dtype)
+
             init_image, mask_image = self._make_images(
                 init_img,
                 init_mask,
@@ -627,9 +633,8 @@ class Generate:
         except RuntimeError:
             # Clear the CUDA cache on an exception
             self.clear_cuda_cache()
-
-            print(traceback.format_exc(), file=sys.stderr)
-            print(">> Could not generate image.")
+            print("** Could not generate image.")
+            raise
 
         toc = time.time()
         print("\n>> Usage stats:")
@@ -909,12 +914,8 @@ class Generate:
         return self._load_generator(".omnibus", "Omnibus")
 
     def _load_generator(self, module, class_name):
-        if self.is_legacy_model(self.model_name):
-            mn = f"ldm.invoke.ckpt_generator{module}"
-            cn = f"Ckpt{class_name}"
-        else:
-            mn = f"ldm.invoke.generator{module}"
-            cn = class_name
+        mn = f"ldm.invoke.generator{module}"
+        cn = class_name
         module = importlib.import_module(mn)
         constructor = getattr(module, cn)
         return constructor(self.model, self.precision)
@@ -976,7 +977,7 @@ class Generate:
         self.generators = {}
 
         seed_everything(random.randrange(0, np.iinfo(np.uint32).max))
-        if self.embedding_path is not None:
+        if self.embedding_path and not model_data.get("ti_embeddings_loaded"):
             print(f'>> Loading embeddings from {self.embedding_path}')
             for root, _, files in os.walk(self.embedding_path):
                 for name in files:
@@ -984,14 +985,24 @@ class Generate:
                     self.model.textual_inversion_manager.load_textual_inversion(
                         ti_path, defer_injecting_tokens=True
                     )
-            print(
-                f'>> Textual inversion triggers: {", ".join(sorted(self.model.textual_inversion_manager.get_all_trigger_strings()))}'
-            )
+            model_data["ti_embeddings_loaded"] = True
+        print(
+            f'>> Textual inversion triggers: {", ".join(sorted(self.model.textual_inversion_manager.get_all_trigger_strings()))}'
+        )
 
         self.model_name = model_name
         self._set_sampler()  # requires self.model_name to be set first
+        self._save_last_used_model(model_name)
         return self.model
 
+    def _save_last_used_model(self,model_name:str):
+        """
+        Save name of the last model used.
+        """
+        model_file_path = Path(Globals.root,'.last_model')
+        with open(model_file_path,'w') as f:
+            f.write(model_name)
+        
     def load_huggingface_concepts(self, concepts: list[str]):
         self.model.textual_inversion_manager.load_huggingface_concepts(concepts)
 
@@ -1032,6 +1043,8 @@ class Generate:
         image_callback=None,
         prefix=None,
     ):
+
+        results = []
         for r in image_list:
             image, seed = r
             try:
@@ -1085,6 +1098,10 @@ class Generate:
             else:
                 r[0] = image
 
+            results.append([image, seed])
+
+        return results
+
     def apply_textmask(
         self, image_path: str, prompt: str, callback, threshold: float = 0.5
     ):
@@ -1113,9 +1130,6 @@ class Generate:
     def sample_to_lowres_estimated_image(self, samples):
         return self._make_base().sample_to_lowres_estimated_image(samples)
 
-    def is_legacy_model(self, model_name) -> bool:
-        return self.model_manager.is_legacy(model_name)
-
     def _set_sampler(self):
         if isinstance(self.model, DiffusionPipeline):
             return self._set_scheduler()

ldm/invoke/CLI.py

@@ -64,7 +64,7 @@ def main():
     Globals.internet_available = args.internet_available and check_internet()
     Globals.disable_xformers = not args.xformers
     Globals.sequential_guidance = args.sequential_guidance
-    Globals.ckpt_convert = args.ckpt_convert
+    Globals.ckpt_convert = True  # always true as of 2.3.4 for LoRA support
 
     # run any post-install patches needed
     run_patches()
@@ -110,6 +110,9 @@ def main():
     else:
         embedding_path = None
 
+    if opt.lora_path:
+        Globals.lora_models_dir = opt.lora_path
+
     # migrate legacy models
     ModelManager.migrate_models()
 
@@ -126,11 +129,13 @@ def main():
             print(f"{e}. Aborting.")
             sys.exit(-1)
 
+    model = opt.model or retrieve_last_used_model()
+
     # creating a Generate object:
     try:
         gen = Generate(
             conf=opt.conf,
-            model=opt.model,
+            model=model,
             sampler_name=opt.sampler_name,
             embedding_path=embedding_path,
             full_precision=opt.full_precision,
@@ -166,14 +171,13 @@ def main():
     if path := opt.autoimport:
         gen.model_manager.heuristic_import(
             str(path),
-            convert=False,
             commit_to_conf=opt.conf,
             config_file_callback=lambda x: _pick_configuration_file(completer,x),
         )
 
     if path := opt.autoconvert:
         gen.model_manager.heuristic_import(
-            str(path), convert=True, commit_to_conf=opt.conf
+            str(path), commit_to_conf=opt.conf
         )
 
     # web server loops forever
@@ -635,7 +639,7 @@ def set_default_output_dir(opt: Args, completer: Completer):
     completer.set_default_dir(opt.outdir)
 
 
-def import_model(model_path: str, gen, opt, completer, convert=False):
+def import_model(model_path: str, gen, opt, completer):
     """
     model_path can be (1) a URL to a .ckpt file; (2) a local .ckpt file path;
     (3) a huggingface repository id; or (4) a local directory containing a
@@ -666,7 +670,6 @@ def import_model(model_path: str, gen, opt, completer, convert=False):
         model_path,
         model_name=model_name,
         description=model_desc,
-        convert=convert,
         config_file_callback=lambda x: _pick_configuration_file(completer,x),
     )
     if not imported_name:
@@ -772,14 +775,10 @@ def convert_model(model_name_or_path: Union[Path, str], gen, opt, completer):
             original_config_file = Path(model_info["config"])
             model_name = model_name_or_path
             model_description = model_info["description"]
-            vae = model_info["vae"]
+            vae_path = model_info.get("vae")
         else:
             print(f"** {model_name_or_path} is not a legacy .ckpt weights file")
             return
-        if vae_repo := ldm.invoke.model_manager.VAE_TO_REPO_ID.get(Path(vae).stem):
-            vae_repo = dict(repo_id=vae_repo)
-        else:
-            vae_repo = None
         model_name = manager.convert_and_import(
             ckpt_path,
             diffusers_path=Path(
@@ -788,11 +787,11 @@ def convert_model(model_name_or_path: Union[Path, str], gen, opt, completer):
             model_name=model_name,
             model_description=model_description,
             original_config_file=original_config_file,
-            vae=vae_repo,
+            vae_path=vae_path,
         )
     else:
         try:
-            import_model(model_name_or_path, gen, opt, completer, convert=True)
+            import_model(model_name_or_path, gen, opt, completer)
         except KeyboardInterrupt:
             return
 
@@ -834,6 +833,7 @@ def edit_model(model_name: str, gen, opt, completer):
     print(f"\n>> Editing model {model_name} from configuration file {opt.conf}")
     new_name = _get_model_name(manager.list_models(), completer, model_name)
 
+    completer.complete_extensions(('.yaml','.ckpt','.safetensors','.pt'))
     for attribute in info.keys():
         if type(info[attribute]) != str:
             continue
@@ -841,6 +841,7 @@ def edit_model(model_name: str, gen, opt, completer):
             continue
         completer.set_line(info[attribute])
         info[attribute] = input(f"{attribute}: ") or info[attribute]
+    completer.complete_extensions(None)
 
     if info["format"] == "diffusers":
         vae = info.get("vae", dict(repo_id=None, path=None, subfolder=None))
@@ -1287,6 +1288,17 @@ def check_internet() -> bool:
     except:
         return False
 
+
+def retrieve_last_used_model()->str:
+    """
+    Return name of the last model used.
+    """
+    model_file_path = Path(Globals.root,'.last_model')
+    if not model_file_path.exists():
+        return None
+    with open(model_file_path,'r') as f:
+        return f.readline()
+
 # This routine performs any patch-ups needed after installation
 def run_patches():
     install_missing_config_files()
@@ -1326,11 +1338,14 @@ def do_version_update(root_version: version.Version, app_version: Union[str, ver
     from release to release. This is not an elegant solution. Instead, the 
     launcher should be moved into the source tree and installed using pip.
     """
+    if root_version < version.Version('v2.3.4'):
+        dest = Path(Globals.root,'loras')
+        dest.mkdir(exist_ok=True)
     if root_version < version.Version('v2.3.3'):
         if sys.platform == "linux":
             print('>> Downloading new version of launcher script and its config file')
             from ldm.util import download_with_progress_bar
-            url_base = f'https://raw.githubusercontent.com/invoke-ai/InvokeAI/release/v{str(app_version)}/installer/templates/'
+            url_base = f'https://raw.githubusercontent.com/invoke-ai/InvokeAI/v{str(app_version)}/installer/templates/'
 
             dest = Path(Globals.root,'invoke.sh.in')
             assert download_with_progress_bar(url_base+'invoke.sh.in',dest)

ldm/invoke/_version.py

@@ -1,2 +1 @@
-
-__version__='2.3.3-rc2'
+__version__='2.3.5-rc1'

ldm/invoke/args.py

@@ -522,8 +522,8 @@ class Args(object):
             '--ckpt_convert',
             action=argparse.BooleanOptionalAction,
             dest='ckpt_convert',
-            default=False,
-            help='Load legacy ckpt files as diffusers. Pass --no-ckpt-convert to inhibit this behavior',
+            default=True,
+            help='Deprecated option. Legacy ckpt files are now always converted to diffusers when loaded.'
         )
         model_group.add_argument(
             '--internet',
@@ -654,6 +654,13 @@ class Args(object):
             type=str,
             help='Path to a directory containing .bin and/or .pt files, or a single .bin/.pt file. You may use subdirectories. (default is ROOTDIR/embeddings)'
         )
+        render_group.add_argument(
+            '--lora_directory',
+            dest='lora_path',
+            default='loras',
+            type=str,
+            help='Path to a directory containing LoRA files; subdirectories are not supported. (default is ROOTDIR/loras)'
+        )
         render_group.add_argument(
             '--embeddings',
             action=argparse.BooleanOptionalAction,
@@ -791,12 +798,12 @@ class Args(object):
             *Model manipulation*
             !models                                   -- list models in configs/models.yaml
             !switch <model_name>                      -- switch to model named <model_name>
-            !import_model /path/to/weights/file.ckpt  -- adds a .ckpt model to your config
+            !import_model /path/to/weights/file       -- imports a model from a ckpt or safetensors file
             !import_model /path/to/weights/           -- interactively import models from a directory
-            !import_model http://path_to_model.ckpt   -- downloads and adds a .ckpt model to your config
-            !import_model hakurei/waifu-diffusion     -- downloads and adds a diffusers model to your config
-            !optimize_model <model_name>              -- converts a .ckpt model to a diffusers model
-            !convert_model /path/to/weights/file.ckpt -- converts a .ckpt file path to a diffusers model
+            !import_model http://path_to_model        -- downloads and adds a ckpt or safetensors model to your config
+            !import_model hakurei/waifu-diffusion     -- downloads and adds a diffusers model to your config using its repo_id
+            !optimize_model <model_name>              -- converts a .ckpt/.safetensors model to a diffusers model
+            !convert_model /path/to/weights/file      -- converts a .ckpt file path to a diffusers model and adds to your config
             !edit_model <model_name>                  -- edit a model's description
             !del_model <model_name>                   -- delete a model
             """

ldm/invoke/ckpt_generator/__init__.py

@@ -1,4 +0,0 @@
-'''
-Initialization file for the ldm.invoke.generator package
-'''
-from .base import CkptGenerator

ldm/invoke/ckpt_generator/base.py

@@ -1,335 +0,0 @@
-'''
-Base class for ldm.invoke.ckpt_generator.*
-including img2img, txt2img, and inpaint
-
-THESE MODULES ARE TRANSITIONAL AND WILL BE REMOVED AT A FUTURE DATE
-WHEN LEGACY CKPT MODEL SUPPORT IS DISCONTINUED.
-'''
-import torch
-import numpy as  np
-import random
-import os
-import os.path as osp
-import traceback
-from tqdm import tqdm, trange
-from PIL import Image, ImageFilter, ImageChops
-import cv2 as cv
-from einops import rearrange, repeat
-from pathlib import Path
-from pytorch_lightning import seed_everything
-import invokeai.assets.web as web_assets
-from ldm.invoke.devices import choose_autocast
-from ldm.models.diffusion.cross_attention_map_saving import AttentionMapSaver
-from ldm.util import rand_perlin_2d
-
-downsampling = 8
-CAUTION_IMG = 'caution.png'
-
-class CkptGenerator():
-    def __init__(self, model, precision):
-        self.model = model
-        self.precision = precision
-        self.seed = None
-        self.latent_channels = model.channels
-        self.downsampling_factor = downsampling   # BUG: should come from model or config
-        self.safety_checker = None
-        self.perlin = 0.0
-        self.threshold = 0
-        self.variation_amount = 0
-        self.with_variations = []
-        self.use_mps_noise = False
-        self.free_gpu_mem = None
-        self.caution_img = None
-
-    # this is going to be overridden in img2img.py, txt2img.py and inpaint.py
-    def get_make_image(self,prompt,**kwargs):
-        """
-        Returns a function returning an image derived from the prompt and the initial image
-        Return value depends on the seed at the time you call it
-        """
-        raise NotImplementedError("image_iterator() must be implemented in a descendent class")
-
-    def set_variation(self, seed, variation_amount, with_variations):
-        self.seed             = seed
-        self.variation_amount = variation_amount
-        self.with_variations  = with_variations
-
-    def generate(self,prompt,init_image,width,height,sampler, iterations=1,seed=None,
-                 image_callback=None, step_callback=None, threshold=0.0, perlin=0.0,
-                 safety_checker:dict=None,
-                 attention_maps_callback = None,
-                 free_gpu_mem: bool=False,
-                 **kwargs):
-        scope = choose_autocast(self.precision)
-        self.safety_checker = safety_checker
-        self.free_gpu_mem = free_gpu_mem
-        attention_maps_images = []
-        attention_maps_callback = lambda saver: attention_maps_images.append(saver.get_stacked_maps_image())
-        make_image = self.get_make_image(
-            prompt,
-            sampler = sampler,
-            init_image    = init_image,
-            width         = width,
-            height        = height,
-            step_callback = step_callback,
-            threshold     = threshold,
-            perlin        = perlin,
-            attention_maps_callback = attention_maps_callback,
-            **kwargs
-        )
-        results             = []
-        seed                = seed if seed is not None and seed >= 0 else self.new_seed()
-        first_seed          = seed
-        seed, initial_noise = self.generate_initial_noise(seed, width, height)
-
-        # There used to be an additional self.model.ema_scope() here, but it breaks
-        # the inpaint-1.5 model. Not sure what it did.... ?
-        with scope(self.model.device.type):
-            for n in trange(iterations, desc='Generating'):
-                x_T = None
-                if self.variation_amount > 0:
-                    seed_everything(seed)
-                    target_noise = self.get_noise(width,height)
-                    x_T = self.slerp(self.variation_amount, initial_noise, target_noise)
-                elif initial_noise is not None:
-                    # i.e. we specified particular variations
-                    x_T = initial_noise
-                else:
-                    seed_everything(seed)
-                    try:
-                        x_T = self.get_noise(width,height)
-                    except:
-                        print('** An error occurred while getting initial noise **')
-                        print(traceback.format_exc())
-
-                image = make_image(x_T)
-
-                if self.safety_checker is not None:
-                    image = self.safety_check(image)
-
-                results.append([image, seed])
-
-                if image_callback is not None:
-                    attention_maps_image = None if len(attention_maps_images)==0 else attention_maps_images[-1]
-                    image_callback(image, seed, first_seed=first_seed, attention_maps_image=attention_maps_image)
-
-                seed = self.new_seed()
-
-        return results
-
-    def sample_to_image(self,samples)->Image.Image:
-        """
-        Given samples returned from a sampler, converts
-        it into a PIL Image
-        """
-        x_samples = self.model.decode_first_stage(samples)
-        x_samples = torch.clamp((x_samples + 1.0) / 2.0, min=0.0, max=1.0)
-        if len(x_samples) != 1:
-            raise Exception(
-                f'>> expected to get a single image, but got {len(x_samples)}')
-        x_sample = 255.0 * rearrange(
-            x_samples[0].cpu().numpy(), 'c h w -> h w c'
-        )
-        return Image.fromarray(x_sample.astype(np.uint8))
-
-        # write an approximate RGB image from latent samples for a single step to PNG
-
-    def repaste_and_color_correct(self, result: Image.Image, init_image: Image.Image, init_mask: Image.Image, mask_blur_radius: int = 8) -> Image.Image:
-        if init_image is None or init_mask is None:
-            return result
-
-        # Get the original alpha channel of the mask if there is one.
-        # Otherwise it is some other black/white image format ('1', 'L' or 'RGB')
-        pil_init_mask = init_mask.getchannel('A') if init_mask.mode == 'RGBA' else init_mask.convert('L')
-        pil_init_image = init_image.convert('RGBA') # Add an alpha channel if one doesn't exist
-
-        # Build an image with only visible pixels from source to use as reference for color-matching.
-        init_rgb_pixels = np.asarray(init_image.convert('RGB'), dtype=np.uint8)
-        init_a_pixels = np.asarray(pil_init_image.getchannel('A'), dtype=np.uint8)
-        init_mask_pixels = np.asarray(pil_init_mask, dtype=np.uint8)
-
-        # Get numpy version of result
-        np_image = np.asarray(result, dtype=np.uint8)
-
-        # Mask and calculate mean and standard deviation
-        mask_pixels = init_a_pixels * init_mask_pixels > 0
-        np_init_rgb_pixels_masked = init_rgb_pixels[mask_pixels, :]
-        np_image_masked = np_image[mask_pixels, :]
-
-        if np_init_rgb_pixels_masked.size > 0:
-            init_means = np_init_rgb_pixels_masked.mean(axis=0)
-            init_std = np_init_rgb_pixels_masked.std(axis=0)
-            gen_means = np_image_masked.mean(axis=0)
-            gen_std = np_image_masked.std(axis=0)
-
-            # Color correct
-            np_matched_result = np_image.copy()
-            np_matched_result[:,:,:] = (((np_matched_result[:,:,:].astype(np.float32) - gen_means[None,None,:]) / gen_std[None,None,:]) * init_std[None,None,:] + init_means[None,None,:]).clip(0, 255).astype(np.uint8)
-            matched_result = Image.fromarray(np_matched_result, mode='RGB')
-        else:
-            matched_result = Image.fromarray(np_image, mode='RGB')
-
-        # Blur the mask out (into init image) by specified amount
-        if mask_blur_radius > 0:
-            nm = np.asarray(pil_init_mask, dtype=np.uint8)
-            nmd = cv.erode(nm, kernel=np.ones((3,3), dtype=np.uint8), iterations=int(mask_blur_radius / 2))
-            pmd = Image.fromarray(nmd, mode='L')
-            blurred_init_mask = pmd.filter(ImageFilter.BoxBlur(mask_blur_radius))
-        else:
-            blurred_init_mask = pil_init_mask
-
-        multiplied_blurred_init_mask = ImageChops.multiply(blurred_init_mask, self.pil_image.split()[-1])
-
-        # Paste original on color-corrected generation (using blurred mask)
-        matched_result.paste(init_image, (0,0), mask = multiplied_blurred_init_mask)
-        return matched_result
-
-
-
-    def sample_to_lowres_estimated_image(self,samples):
-        # origingally adapted from code by @erucipe and @keturn here:
-        # https://discuss.huggingface.co/t/decoding-latents-to-rgb-without-upscaling/23204/7
-
-        # these updated numbers for v1.5 are from @torridgristle
-        v1_5_latent_rgb_factors = torch.tensor([
-            #    R        G        B
-            [ 0.3444,  0.1385,  0.0670], # L1
-            [ 0.1247,  0.4027,  0.1494], # L2
-            [-0.3192,  0.2513,  0.2103], # L3
-            [-0.1307, -0.1874, -0.7445]  # L4
-        ], dtype=samples.dtype, device=samples.device)
-
-        latent_image = samples[0].permute(1, 2, 0) @ v1_5_latent_rgb_factors
-        latents_ubyte = (((latent_image + 1) / 2)
-                         .clamp(0, 1)  # change scale from -1..1 to 0..1
-                         .mul(0xFF)  # to 0..255
-                         .byte()).cpu()
-
-        return Image.fromarray(latents_ubyte.numpy())
-
-    def generate_initial_noise(self, seed, width, height):
-        initial_noise = None
-        if self.variation_amount > 0 or len(self.with_variations) > 0:
-            # use fixed initial noise plus random noise per iteration
-            seed_everything(seed)
-            initial_noise = self.get_noise(width,height)
-            for v_seed, v_weight in self.with_variations:
-                seed = v_seed
-                seed_everything(seed)
-                next_noise = self.get_noise(width,height)
-                initial_noise = self.slerp(v_weight, initial_noise, next_noise)
-            if self.variation_amount > 0:
-                random.seed() # reset RNG to an actually random state, so we can get a random seed for variations
-                seed = random.randrange(0,np.iinfo(np.uint32).max)
-            return (seed, initial_noise)
-        else:
-            return (seed, None)
-
-    # returns a tensor filled with random numbers from a normal distribution
-    def get_noise(self,width,height):
-        """
-        Returns a tensor filled with random numbers, either form a normal distribution
-        (txt2img) or from the latent image (img2img, inpaint)
-        """
-        raise NotImplementedError("get_noise() must be implemented in a descendent class")
-
-    def get_perlin_noise(self,width,height):
-        fixdevice = 'cpu' if (self.model.device.type == 'mps') else self.model.device
-        return torch.stack([rand_perlin_2d((height, width), (8, 8), device = self.model.device).to(fixdevice) for _ in range(self.latent_channels)], dim=0).to(self.model.device)
-
-    def new_seed(self):
-        self.seed = random.randrange(0, np.iinfo(np.uint32).max)
-        return self.seed
-
-    def slerp(self, t, v0, v1, DOT_THRESHOLD=0.9995):
-        '''
-        Spherical linear interpolation
-        Args:
-            t (float/np.ndarray): Float value between 0.0 and 1.0
-            v0 (np.ndarray): Starting vector
-            v1 (np.ndarray): Final vector
-            DOT_THRESHOLD (float): Threshold for considering the two vectors as
-                                colineal. Not recommended to alter this.
-        Returns:
-            v2 (np.ndarray): Interpolation vector between v0 and v1
-        '''
-        inputs_are_torch = False
-        if not isinstance(v0, np.ndarray):
-            inputs_are_torch = True
-            v0 = v0.detach().cpu().numpy()
-        if not isinstance(v1, np.ndarray):
-            inputs_are_torch = True
-            v1 = v1.detach().cpu().numpy()
-
-        dot = np.sum(v0 * v1 / (np.linalg.norm(v0) * np.linalg.norm(v1)))
-        if np.abs(dot) > DOT_THRESHOLD:
-            v2 = (1 - t) * v0 + t * v1
-        else:
-            theta_0 = np.arccos(dot)
-            sin_theta_0 = np.sin(theta_0)
-            theta_t = theta_0 * t
-            sin_theta_t = np.sin(theta_t)
-            s0 = np.sin(theta_0 - theta_t) / sin_theta_0
-            s1 = sin_theta_t / sin_theta_0
-            v2 = s0 * v0 + s1 * v1
-
-        if inputs_are_torch:
-            v2 = torch.from_numpy(v2).to(self.model.device)
-
-        return v2
-
-    def safety_check(self,image:Image.Image):
-        '''
-        If the CompViz safety checker flags an NSFW image, we
-        blur it out.
-        '''
-        import diffusers
-
-        checker = self.safety_checker['checker']
-        extractor = self.safety_checker['extractor']
-        features = extractor([image], return_tensors="pt")
-        features.to(self.model.device)
-
-        # unfortunately checker requires the numpy version, so we have to convert back
-        x_image = np.array(image).astype(np.float32) / 255.0
-        x_image = x_image[None].transpose(0, 3, 1, 2)
-
-        diffusers.logging.set_verbosity_error()
-        checked_image, has_nsfw_concept = checker(images=x_image, clip_input=features.pixel_values)
-        if has_nsfw_concept[0]:
-            print('** An image with potential non-safe content has been detected. A blurred image will be returned. **')
-            return self.blur(image)
-        else:
-            return image
-
-    def blur(self,input):
-        blurry = input.filter(filter=ImageFilter.GaussianBlur(radius=32))
-        try:
-            caution = self.get_caution_img()
-            if caution:
-                blurry.paste(caution,(0,0),caution)
-        except FileNotFoundError:
-            pass
-        return blurry
-
-    def get_caution_img(self):
-        path = None
-        if self.caution_img:
-            return self.caution_img
-        path = Path(web_assets.__path__[0]) / CAUTION_IMG
-        caution = Image.open(path)
-        self.caution_img = caution.resize((caution.width // 2, caution.height //2))
-        return self.caution_img
-
-    # this is a handy routine for debugging use. Given a generated sample,
-    # convert it into a PNG image and store it at the indicated path
-    def save_sample(self, sample, filepath):
-        image = self.sample_to_image(sample)
-        dirname = os.path.dirname(filepath) or '.'
-        if not os.path.exists(dirname):
-            print(f'** creating directory {dirname}')
-            os.makedirs(dirname, exist_ok=True)
-        image.save(filepath,'PNG')
-
-    def torch_dtype(self)->torch.dtype:
-        return torch.float16 if self.precision == 'float16' else torch.float32

ldm/invoke/ckpt_generator/embiggen.py

@@ -1,501 +0,0 @@
-'''
-ldm.invoke.ckpt_generator.embiggen descends from ldm.invoke.ckpt_generator
-and generates with ldm.invoke.ckpt_generator.img2img
-'''
-
-import torch
-import numpy as  np
-from tqdm import trange
-from PIL               import Image
-from ldm.invoke.ckpt_generator.base      import CkptGenerator
-from ldm.invoke.ckpt_generator.img2img   import CkptImg2Img
-from ldm.invoke.devices import choose_autocast
-from ldm.models.diffusion.ddim     import DDIMSampler
-
-class CkptEmbiggen(CkptGenerator):
-    def __init__(self, model, precision):
-        super().__init__(model, precision)
-        self.init_latent         = None
-
-    # Replace generate because Embiggen doesn't need/use most of what it does normallly
-    def generate(self,prompt,iterations=1,seed=None,
-                 image_callback=None, step_callback=None,
-                 **kwargs):
-
-        scope      = choose_autocast(self.precision)
-        make_image = self.get_make_image(
-            prompt,
-            step_callback = step_callback,
-            **kwargs
-        )
-        results             = []
-        seed                = seed if seed else self.new_seed()
-
-        # Noise will be generated by the Img2Img generator when called
-        with scope(self.model.device.type), self.model.ema_scope():
-            for n in trange(iterations, desc='Generating'):
-                # make_image will call Img2Img which will do the equivalent of get_noise itself
-                image = make_image()
-                results.append([image, seed])
-                if image_callback is not None:
-                    image_callback(image, seed, prompt_in=prompt)
-                seed = self.new_seed()
-        return results
-
-    @torch.no_grad()
-    def get_make_image(
-        self,
-        prompt,
-        sampler,
-        steps,
-        cfg_scale,
-        ddim_eta,
-        conditioning,
-        init_img,
-        strength,
-        width,
-        height,
-        embiggen,
-        embiggen_tiles,
-        step_callback=None,
-        **kwargs
-    ):
-        """
-        Returns a function returning an image derived from the prompt and multi-stage twice-baked potato layering over the img2img on the initial image
-        Return value depends on the seed at the time you call it
-        """
-        assert not sampler.uses_inpainting_model(), "--embiggen is not supported by inpainting models"
-
-        # Construct embiggen arg array, and sanity check arguments
-        if embiggen == None:  # embiggen can also be called with just embiggen_tiles
-            embiggen = [1.0]  # If not specified, assume no scaling
-        elif embiggen[0] < 0:
-            embiggen[0] = 1.0
-            print(
-                '>> Embiggen scaling factor cannot be negative, fell back to the default of 1.0 !')
-        if len(embiggen) < 2:
-            embiggen.append(0.75)
-        elif embiggen[1] > 1.0 or embiggen[1] < 0:
-            embiggen[1] = 0.75
-            print('>> Embiggen upscaling strength for ESRGAN must be between 0 and 1, fell back to the default of 0.75 !')
-        if len(embiggen) < 3:
-            embiggen.append(0.25)
-        elif embiggen[2] < 0:
-            embiggen[2] = 0.25
-            print('>> Overlap size for Embiggen must be a positive ratio between 0 and 1 OR a number of pixels, fell back to the default of 0.25 !')
-
-        # Convert tiles from their user-freindly count-from-one to count-from-zero, because we need to do modulo math
-        # and then sort them, because... people.
-        if embiggen_tiles:
-            embiggen_tiles = list(map(lambda n: n-1, embiggen_tiles))
-            embiggen_tiles.sort()
-
-        if strength >= 0.5:
-            print(f'* WARNING: Embiggen may produce mirror motifs if the strength (-f) is too high (currently {strength}). Try values between 0.35-0.45.')
-
-        # Prep img2img generator, since we wrap over it
-        gen_img2img = CkptImg2Img(self.model,self.precision)
-
-        # Open original init image (not a tensor) to manipulate
-        initsuperimage = Image.open(init_img)
-
-        with Image.open(init_img) as img:
-            initsuperimage = img.convert('RGB')
-
-        # Size of the target super init image in pixels
-        initsuperwidth, initsuperheight = initsuperimage.size
-
-        # Increase by scaling factor if not already resized, using ESRGAN as able
-        if embiggen[0] != 1.0:
-            initsuperwidth = round(initsuperwidth*embiggen[0])
-            initsuperheight = round(initsuperheight*embiggen[0])
-            if embiggen[1] > 0:  # No point in ESRGAN upscaling if strength is set zero
-                from ldm.invoke.restoration.realesrgan import ESRGAN
-                esrgan = ESRGAN()
-                print(
-                    f'>> ESRGAN upscaling init image prior to cutting with Embiggen with strength {embiggen[1]}')
-                if embiggen[0] > 2:
-                    initsuperimage = esrgan.process(
-                        initsuperimage,
-                        embiggen[1],  # upscale strength
-                        self.seed,
-                        4,  # upscale scale
-                    )
-                else:
-                    initsuperimage = esrgan.process(
-                        initsuperimage,
-                        embiggen[1],  # upscale strength
-                        self.seed,
-                        2,  # upscale scale
-                    )
-            # We could keep recursively re-running ESRGAN for a requested embiggen[0] larger than 4x
-            #   but from personal experiance it doesn't greatly improve anything after 4x
-            # Resize to target scaling factor resolution
-            initsuperimage = initsuperimage.resize(
-                (initsuperwidth, initsuperheight), Image.Resampling.LANCZOS)
-
-        # Use width and height as tile widths and height
-        # Determine buffer size in pixels
-        if embiggen[2] < 1:
-            if embiggen[2] < 0:
-                embiggen[2] = 0
-            overlap_size_x = round(embiggen[2] * width)
-            overlap_size_y = round(embiggen[2] * height)
-        else:
-            overlap_size_x = round(embiggen[2])
-            overlap_size_y = round(embiggen[2])
-
-        # With overall image width and height known, determine how many tiles we need
-        def ceildiv(a, b):
-            return -1 * (-a // b)
-
-        # X and Y needs to be determined independantly (we may have savings on one based on the buffer pixel count)
-        # (initsuperwidth - width) is the area remaining to the right that we need to layers tiles to fill
-        # (width - overlap_size_x) is how much new we can fill with a single tile
-        emb_tiles_x = 1
-        emb_tiles_y = 1
-        if (initsuperwidth - width) > 0:
-            emb_tiles_x = ceildiv(initsuperwidth - width,
-                                  width - overlap_size_x) + 1
-        if (initsuperheight - height) > 0:
-            emb_tiles_y = ceildiv(initsuperheight - height,
-                                  height - overlap_size_y) + 1
-        # Sanity
-        assert emb_tiles_x > 1 or emb_tiles_y > 1, f'ERROR: Based on the requested dimensions of {initsuperwidth}x{initsuperheight} and tiles of {width}x{height} you don\'t need to Embiggen! Check your arguments.'
-
-        # Prep alpha layers --------------
-        # https://stackoverflow.com/questions/69321734/how-to-create-different-transparency-like-gradient-with-python-pil
-        # agradientL is Left-side transparent
-        agradientL = Image.linear_gradient('L').rotate(
-            90).resize((overlap_size_x, height))
-        # agradientT is Top-side transparent
-        agradientT = Image.linear_gradient('L').resize((width, overlap_size_y))
-        # radial corner is the left-top corner, made full circle then cut to just the left-top quadrant
-        agradientC = Image.new('L', (256, 256))
-        for y in range(256):
-            for x in range(256):
-                # Find distance to lower right corner (numpy takes arrays)
-                distanceToLR = np.sqrt([(255 - x) ** 2 + (255 - y) ** 2])[0]
-                # Clamp values to max 255
-                if distanceToLR > 255:
-                    distanceToLR = 255
-                #Place the pixel as invert of distance
-                agradientC.putpixel((x, y), round(255 - distanceToLR))
-
-        # Create alternative asymmetric diagonal corner to use on "tailing" intersections to prevent hard edges
-        # Fits for a left-fading gradient on the bottom side and full opacity on the right side.
-        agradientAsymC = Image.new('L', (256, 256))
-        for y in range(256):
-            for x in range(256):
-                value = round(max(0, x-(255-y)) * (255 / max(1,y)))
-                #Clamp values
-                value = max(0, value)
-                value = min(255, value)
-                agradientAsymC.putpixel((x, y), value)
-
-        # Create alpha layers default fully white
-        alphaLayerL = Image.new("L", (width, height), 255)
-        alphaLayerT = Image.new("L", (width, height), 255)
-        alphaLayerLTC = Image.new("L", (width, height), 255)
-        # Paste gradients into alpha layers
-        alphaLayerL.paste(agradientL, (0, 0))
-        alphaLayerT.paste(agradientT, (0, 0))
-        alphaLayerLTC.paste(agradientL, (0, 0))
-        alphaLayerLTC.paste(agradientT, (0, 0))
-        alphaLayerLTC.paste(agradientC.resize((overlap_size_x, overlap_size_y)), (0, 0))
-        # make masks with an asymmetric upper-right corner so when the curved transparent corner of the next tile
-        # to its right is placed it doesn't reveal a hard trailing semi-transparent edge in the overlapping space
-        alphaLayerTaC = alphaLayerT.copy()
-        alphaLayerTaC.paste(agradientAsymC.rotate(270).resize((overlap_size_x, overlap_size_y)), (width - overlap_size_x, 0))
-        alphaLayerLTaC = alphaLayerLTC.copy()
-        alphaLayerLTaC.paste(agradientAsymC.rotate(270).resize((overlap_size_x, overlap_size_y)), (width - overlap_size_x, 0))
-
-        if embiggen_tiles:
-            # Individual unconnected sides
-            alphaLayerR = Image.new("L", (width, height), 255)
-            alphaLayerR.paste(agradientL.rotate(
-                180), (width - overlap_size_x, 0))
-            alphaLayerB = Image.new("L", (width, height), 255)
-            alphaLayerB.paste(agradientT.rotate(
-                180), (0, height - overlap_size_y))
-            alphaLayerTB = Image.new("L", (width, height), 255)
-            alphaLayerTB.paste(agradientT, (0, 0))
-            alphaLayerTB.paste(agradientT.rotate(
-                180), (0, height - overlap_size_y))
-            alphaLayerLR = Image.new("L", (width, height), 255)
-            alphaLayerLR.paste(agradientL, (0, 0))
-            alphaLayerLR.paste(agradientL.rotate(
-                180), (width - overlap_size_x, 0))
-
-            # Sides and corner Layers
-            alphaLayerRBC = Image.new("L", (width, height), 255)
-            alphaLayerRBC.paste(agradientL.rotate(
-                180), (width - overlap_size_x, 0))
-            alphaLayerRBC.paste(agradientT.rotate(
-                180), (0, height - overlap_size_y))
-            alphaLayerRBC.paste(agradientC.rotate(180).resize(
-                (overlap_size_x, overlap_size_y)), (width - overlap_size_x, height - overlap_size_y))
-            alphaLayerLBC = Image.new("L", (width, height), 255)
-            alphaLayerLBC.paste(agradientL, (0, 0))
-            alphaLayerLBC.paste(agradientT.rotate(
-                180), (0, height - overlap_size_y))
-            alphaLayerLBC.paste(agradientC.rotate(90).resize(
-                (overlap_size_x, overlap_size_y)), (0, height - overlap_size_y))
-            alphaLayerRTC = Image.new("L", (width, height), 255)
-            alphaLayerRTC.paste(agradientL.rotate(
-                180), (width - overlap_size_x, 0))
-            alphaLayerRTC.paste(agradientT, (0, 0))
-            alphaLayerRTC.paste(agradientC.rotate(270).resize(
-                (overlap_size_x, overlap_size_y)), (width - overlap_size_x, 0))
-
-            # All but X layers
-            alphaLayerABT = Image.new("L", (width, height), 255)
-            alphaLayerABT.paste(alphaLayerLBC, (0, 0))
-            alphaLayerABT.paste(agradientL.rotate(
-                180), (width - overlap_size_x, 0))
-            alphaLayerABT.paste(agradientC.rotate(180).resize(
-                (overlap_size_x, overlap_size_y)), (width - overlap_size_x, height - overlap_size_y))
-            alphaLayerABL = Image.new("L", (width, height), 255)
-            alphaLayerABL.paste(alphaLayerRTC, (0, 0))
-            alphaLayerABL.paste(agradientT.rotate(
-                180), (0, height - overlap_size_y))
-            alphaLayerABL.paste(agradientC.rotate(180).resize(
-                (overlap_size_x, overlap_size_y)), (width - overlap_size_x, height - overlap_size_y))
-            alphaLayerABR = Image.new("L", (width, height), 255)
-            alphaLayerABR.paste(alphaLayerLBC, (0, 0))
-            alphaLayerABR.paste(agradientT, (0, 0))
-            alphaLayerABR.paste(agradientC.resize(
-                (overlap_size_x, overlap_size_y)), (0, 0))
-            alphaLayerABB = Image.new("L", (width, height), 255)
-            alphaLayerABB.paste(alphaLayerRTC, (0, 0))
-            alphaLayerABB.paste(agradientL, (0, 0))
-            alphaLayerABB.paste(agradientC.resize(
-                (overlap_size_x, overlap_size_y)), (0, 0))
-
-            # All-around layer
-            alphaLayerAA = Image.new("L", (width, height), 255)
-            alphaLayerAA.paste(alphaLayerABT, (0, 0))
-            alphaLayerAA.paste(agradientT, (0, 0))
-            alphaLayerAA.paste(agradientC.resize(
-                (overlap_size_x, overlap_size_y)), (0, 0))
-            alphaLayerAA.paste(agradientC.rotate(270).resize(
-                (overlap_size_x, overlap_size_y)), (width - overlap_size_x, 0))
-
-        # Clean up temporary gradients
-        del agradientL
-        del agradientT
-        del agradientC
-
-        def make_image():
-            # Make main tiles -------------------------------------------------
-            if embiggen_tiles:
-                print(f'>> Making {len(embiggen_tiles)} Embiggen tiles...')
-            else:
-                print(
-                    f'>> Making {(emb_tiles_x * emb_tiles_y)} Embiggen tiles ({emb_tiles_x}x{emb_tiles_y})...')
-
-            emb_tile_store = []
-            # Although we could use the same seed for every tile for determinism, at higher strengths this may
-            # produce duplicated structures for each tile and make the tiling effect more obvious
-            # instead track and iterate a local seed we pass to Img2Img
-            seed = self.seed
-            seedintlimit = np.iinfo(np.uint32).max - 1 # only retreive this one from numpy
-
-            for tile in range(emb_tiles_x * emb_tiles_y):
-                # Don't iterate on first tile
-                if tile != 0:
-                    if seed < seedintlimit:
-                        seed += 1
-                    else:
-                        seed = 0
-
-                # Determine if this is a re-run and replace
-                if embiggen_tiles and not tile in embiggen_tiles:
-                    continue
-                # Get row and column entries
-                emb_row_i = tile // emb_tiles_x
-                emb_column_i = tile % emb_tiles_x
-                # Determine bounds to cut up the init image
-                # Determine upper-left point
-                if emb_column_i + 1 == emb_tiles_x:
-                    left = initsuperwidth - width
-                else:
-                    left = round(emb_column_i * (width - overlap_size_x))
-                if emb_row_i + 1 == emb_tiles_y:
-                    top = initsuperheight - height
-                else:
-                    top = round(emb_row_i * (height - overlap_size_y))
-                right = left + width
-                bottom = top + height
-
-                # Cropped image of above dimension (does not modify the original)
-                newinitimage = initsuperimage.crop((left, top, right, bottom))
-                # DEBUG:
-                # newinitimagepath = init_img[0:-4] + f'_emb_Ti{tile}.png'
-                # newinitimage.save(newinitimagepath)
-
-                if embiggen_tiles:
-                    print(
-                        f'Making tile #{tile + 1} ({embiggen_tiles.index(tile) + 1} of {len(embiggen_tiles)} requested)')
-                else:
-                    print(
-                        f'Starting {tile + 1} of {(emb_tiles_x * emb_tiles_y)} tiles')
-
-                # create a torch tensor from an Image
-                newinitimage = np.array(
-                    newinitimage).astype(np.float32) / 255.0
-                newinitimage = newinitimage[None].transpose(0, 3, 1, 2)
-                newinitimage = torch.from_numpy(newinitimage)
-                newinitimage = 2.0 * newinitimage - 1.0
-                newinitimage = newinitimage.to(self.model.device)
-
-                tile_results = gen_img2img.generate(
-                    prompt,
-                    iterations     = 1,
-                    seed           = seed,
-                    sampler        = DDIMSampler(self.model, device=self.model.device),
-                    steps          = steps,
-                    cfg_scale      = cfg_scale,
-                    conditioning   = conditioning,
-                    ddim_eta       = ddim_eta,
-                    image_callback = None,  # called only after the final image is generated
-                    step_callback  = step_callback,   # called after each intermediate image is generated
-                    width          = width,
-                    height         = height,
-                    init_image     = newinitimage,    # notice that init_image is different from init_img
-                    mask_image     = None,
-                    strength       = strength,
-                )
-
-                emb_tile_store.append(tile_results[0][0])
-                # DEBUG (but, also has other uses), worth saving if you want tiles without a transparency overlap to manually composite
-                # emb_tile_store[-1].save(init_img[0:-4] + f'_emb_To{tile}.png')
-                del newinitimage
-
-            # Sanity check we have them all
-            if len(emb_tile_store) == (emb_tiles_x * emb_tiles_y) or (embiggen_tiles != [] and len(emb_tile_store) == len(embiggen_tiles)):
-                outputsuperimage = Image.new(
-                    "RGBA", (initsuperwidth, initsuperheight))
-                if embiggen_tiles:
-                    outputsuperimage.alpha_composite(
-                        initsuperimage.convert('RGBA'), (0, 0))
-                for tile in range(emb_tiles_x * emb_tiles_y):
-                    if embiggen_tiles:
-                        if tile in embiggen_tiles:
-                            intileimage = emb_tile_store.pop(0)
-                        else:
-                            continue
-                    else:
-                        intileimage = emb_tile_store[tile]
-                    intileimage = intileimage.convert('RGBA')
-                    # Get row and column entries
-                    emb_row_i = tile // emb_tiles_x
-                    emb_column_i = tile % emb_tiles_x
-                    if emb_row_i == 0 and emb_column_i == 0 and not embiggen_tiles:
-                        left = 0
-                        top = 0
-                    else:
-                        # Determine upper-left point
-                        if emb_column_i + 1 == emb_tiles_x:
-                            left = initsuperwidth - width
-                        else:
-                            left = round(emb_column_i *
-                                         (width - overlap_size_x))
-                        if emb_row_i + 1 == emb_tiles_y:
-                            top = initsuperheight - height
-                        else:
-                            top = round(emb_row_i * (height - overlap_size_y))
-                        # Handle gradients for various conditions
-                        # Handle emb_rerun case
-                        if embiggen_tiles:
-                            # top of image
-                            if emb_row_i == 0:
-                                if emb_column_i == 0:
-                                    if (tile+1) in embiggen_tiles:  # Look-ahead right
-                                        if (tile+emb_tiles_x) not in embiggen_tiles:  # Look-ahead down
-                                            intileimage.putalpha(alphaLayerB)
-                                        # Otherwise do nothing on this tile
-                                    elif (tile+emb_tiles_x) in embiggen_tiles:  # Look-ahead down only
-                                        intileimage.putalpha(alphaLayerR)
-                                    else:
-                                        intileimage.putalpha(alphaLayerRBC)
-                                elif emb_column_i == emb_tiles_x - 1:
-                                    if (tile+emb_tiles_x) in embiggen_tiles:  # Look-ahead down
-                                        intileimage.putalpha(alphaLayerL)
-                                    else:
-                                        intileimage.putalpha(alphaLayerLBC)
-                                else:
-                                    if (tile+1) in embiggen_tiles:  # Look-ahead right
-                                        if (tile+emb_tiles_x) in embiggen_tiles:  # Look-ahead down
-                                            intileimage.putalpha(alphaLayerL)
-                                        else:
-                                            intileimage.putalpha(alphaLayerLBC)
-                                    elif (tile+emb_tiles_x) in embiggen_tiles:  # Look-ahead down only
-                                        intileimage.putalpha(alphaLayerLR)
-                                    else:
-                                        intileimage.putalpha(alphaLayerABT)
-                            # bottom of image
-                            elif emb_row_i == emb_tiles_y - 1:
-                                if emb_column_i == 0:
-                                    if (tile+1) in embiggen_tiles: # Look-ahead right
-                                        intileimage.putalpha(alphaLayerTaC)
-                                    else:
-                                        intileimage.putalpha(alphaLayerRTC)
-                                elif emb_column_i == emb_tiles_x - 1:
-                                    # No tiles to look ahead to
-                                    intileimage.putalpha(alphaLayerLTC)
-                                else:
-                                    if (tile+1) in embiggen_tiles: # Look-ahead right
-                                        intileimage.putalpha(alphaLayerLTaC)
-                                    else:
-                                        intileimage.putalpha(alphaLayerABB)
-                            # vertical middle of image
-                            else:
-                                if emb_column_i == 0:
-                                    if (tile+1) in embiggen_tiles: # Look-ahead right
-                                        if (tile+emb_tiles_x) in embiggen_tiles: # Look-ahead down
-                                            intileimage.putalpha(alphaLayerTaC)
-                                        else:
-                                            intileimage.putalpha(alphaLayerTB)
-                                    elif (tile+emb_tiles_x) in embiggen_tiles:  # Look-ahead down only
-                                        intileimage.putalpha(alphaLayerRTC)
-                                    else:
-                                        intileimage.putalpha(alphaLayerABL)
-                                elif emb_column_i == emb_tiles_x - 1:
-                                    if (tile+emb_tiles_x) in embiggen_tiles:  # Look-ahead down
-                                        intileimage.putalpha(alphaLayerLTC)
-                                    else:
-                                        intileimage.putalpha(alphaLayerABR)
-                                else:
-                                    if (tile+1) in embiggen_tiles: # Look-ahead right
-                                        if (tile+emb_tiles_x) in embiggen_tiles: # Look-ahead down
-                                            intileimage.putalpha(alphaLayerLTaC)
-                                        else:
-                                            intileimage.putalpha(alphaLayerABR)
-                                    elif (tile+emb_tiles_x) in embiggen_tiles:  # Look-ahead down only
-                                        intileimage.putalpha(alphaLayerABB)
-                                    else:
-                                        intileimage.putalpha(alphaLayerAA)
-                        # Handle normal tiling case (much simpler - since we tile left to right, top to bottom)
-                        else:
-                            if emb_row_i == 0 and emb_column_i >= 1:
-                                intileimage.putalpha(alphaLayerL)
-                            elif emb_row_i >= 1 and emb_column_i == 0:
-                                if emb_column_i + 1 == emb_tiles_x: # If we don't have anything that can be placed to the right
-                                    intileimage.putalpha(alphaLayerT)
-                                else:
-                                    intileimage.putalpha(alphaLayerTaC)
-                            else:
-                                if emb_column_i + 1 == emb_tiles_x: # If we don't have anything that can be placed to the right
-                                    intileimage.putalpha(alphaLayerLTC)
-                                else:
-                                    intileimage.putalpha(alphaLayerLTaC)
-                    # Layer tile onto final image
-                    outputsuperimage.alpha_composite(intileimage, (left, top))
-            else:
-                print(f'Error: could not find all Embiggen output tiles in memory? Something must have gone wrong with img2img generation.')
-
-            # after internal loops and patching up return Embiggen image
-            return outputsuperimage
-        # end of function declaration
-        return make_image

ldm/invoke/ckpt_generator/img2img.py

@@ -1,97 +0,0 @@
-'''
-ldm.invoke.ckpt_generator.img2img descends from ldm.invoke.generator
-'''
-
-import torch
-import numpy as  np
-import PIL
-from torch import Tensor
-from PIL import Image
-from ldm.invoke.devices import choose_autocast
-from ldm.invoke.ckpt_generator.base import CkptGenerator
-from ldm.models.diffusion.ddim import DDIMSampler
-from ldm.models.diffusion.shared_invokeai_diffusion import InvokeAIDiffuserComponent
-
-class CkptImg2Img(CkptGenerator):
-    def __init__(self, model, precision):
-        super().__init__(model, precision)
-        self.init_latent = None    # by get_noise()
-
-    def get_make_image(self,prompt,sampler,steps,cfg_scale,ddim_eta,
-                       conditioning,init_image,strength,step_callback=None,threshold=0.0,perlin=0.0,**kwargs):
-        """
-        Returns a function returning an image derived from the prompt and the initial image
-        Return value depends on the seed at the time you call it.
-        """
-        self.perlin = perlin
-
-        sampler.make_schedule(
-            ddim_num_steps=steps, ddim_eta=ddim_eta, verbose=False
-        )
-
-        if isinstance(init_image, PIL.Image.Image):
-            init_image = self._image_to_tensor(init_image.convert('RGB'))
-
-        scope = choose_autocast(self.precision)
-        with scope(self.model.device.type):
-            self.init_latent = self.model.get_first_stage_encoding(
-                self.model.encode_first_stage(init_image)
-            ) # move to latent space
-
-        t_enc = int(strength * steps)
-        uc, c, extra_conditioning_info   = conditioning
-
-        def make_image(x_T):
-            # encode (scaled latent)
-            z_enc = sampler.stochastic_encode(
-                self.init_latent,
-                torch.tensor([t_enc - 1]).to(self.model.device),
-                noise=x_T
-            )
-
-            if self.free_gpu_mem and self.model.model.device != self.model.device:
-                self.model.model.to(self.model.device)
-
-            # decode it
-            samples = sampler.decode(
-                z_enc,
-                c,
-                t_enc,
-                img_callback = step_callback,
-                unconditional_guidance_scale=cfg_scale,
-                unconditional_conditioning=uc,
-                init_latent = self.init_latent, # changes how noising is performed in ksampler
-                extra_conditioning_info = extra_conditioning_info,
-                all_timesteps_count = steps
-            )
-
-            if self.free_gpu_mem:
-                self.model.model.to("cpu")
-
-            return self.sample_to_image(samples)
-
-        return make_image
-
-    def get_noise(self,width,height):
-        device      = self.model.device
-        init_latent = self.init_latent
-        assert init_latent is not None,'call to get_noise() when init_latent not set'
-        if device.type == 'mps':
-            x = torch.randn_like(init_latent, device='cpu').to(device)
-        else:
-            x = torch.randn_like(init_latent, device=device)
-        if self.perlin > 0.0:
-            shape = init_latent.shape
-            x = (1-self.perlin)*x + self.perlin*self.get_perlin_noise(shape[3], shape[2])
-        return x
-
-    def _image_to_tensor(self, image:Image, normalize:bool=True)->Tensor:
-        image = np.array(image).astype(np.float32) / 255.0
-        if len(image.shape) == 2:  # 'L' image, as in a mask
-            image = image[None,None]
-        else:                      # 'RGB' image
-            image = image[None].transpose(0, 3, 1, 2)
-        image = torch.from_numpy(image)
-        if normalize:
-            image = 2.0 * image - 1.0
-        return image.to(self.model.device)

ldm/invoke/ckpt_generator/inpaint.py

@@ -1,358 +0,0 @@
-'''
-ldm.invoke.ckpt_generator.inpaint descends from ldm.invoke.ckpt_generator
-'''
-
-import math
-import torch
-import torchvision.transforms as T
-import numpy as  np
-import cv2 as cv
-import PIL
-from PIL import Image, ImageFilter, ImageOps, ImageChops
-from skimage.exposure.histogram_matching import match_histograms
-from einops import rearrange, repeat
-from ldm.invoke.devices             import choose_autocast
-from ldm.invoke.ckpt_generator.img2img   import CkptImg2Img
-from ldm.models.diffusion.ddim     import DDIMSampler
-from ldm.models.diffusion.ksampler import KSampler
-from ldm.invoke.generator.base import downsampling
-from ldm.util import debug_image
-from ldm.invoke.patchmatch import PatchMatch 
-from ldm.invoke.globals import Globals
-
-def infill_methods()->list[str]:
-    methods = list()
-    if PatchMatch.patchmatch_available():
-        methods.append('patchmatch')
-    methods.append('tile')
-    return methods
-
-class CkptInpaint(CkptImg2Img):
-    def __init__(self, model, precision):
-        self.init_latent = None
-        self.pil_image = None
-        self.pil_mask = None
-        self.mask_blur_radius = 0
-        self.infill_method = None
-        super().__init__(model, precision)
-
-    # Outpaint support code
-    def get_tile_images(self, image: np.ndarray, width=8, height=8):
-        _nrows, _ncols, depth = image.shape
-        _strides = image.strides
-
-        nrows, _m = divmod(_nrows, height)
-        ncols, _n = divmod(_ncols, width)
-        if _m != 0 or _n != 0:
-            return None
-
-        return np.lib.stride_tricks.as_strided(
-            np.ravel(image),
-            shape=(nrows, ncols, height, width, depth),
-            strides=(height * _strides[0], width * _strides[1], *_strides),
-            writeable=False
-        )
-
-    def infill_patchmatch(self, im: Image.Image) -> Image:
-        if im.mode != 'RGBA':
-            return im
-
-        # Skip patchmatch if patchmatch isn't available
-        if not PatchMatch.patchmatch_available():
-            return im
-
-        # Patchmatch (note, we may want to expose patch_size? Increasing it significantly impacts performance though)
-        im_patched_np = PatchMatch.inpaint(im.convert('RGB'), ImageOps.invert(im.split()[-1]), patch_size = 3)
-        im_patched = Image.fromarray(im_patched_np, mode = 'RGB')
-        return im_patched
-
-    def tile_fill_missing(self, im: Image.Image, tile_size: int = 16, seed: int = None) -> Image:
-        # Only fill if there's an alpha layer
-        if im.mode != 'RGBA':
-            return im
-
-        a = np.asarray(im, dtype=np.uint8)
-
-        tile_size = (tile_size, tile_size)
-
-        # Get the image as tiles of a specified size
-        tiles = self.get_tile_images(a,*tile_size).copy()
-
-        # Get the mask as tiles
-        tiles_mask = tiles[:,:,:,:,3]
-
-        # Find any mask tiles with any fully transparent pixels (we will be replacing these later)
-        tmask_shape = tiles_mask.shape
-        tiles_mask = tiles_mask.reshape(math.prod(tiles_mask.shape))
-        n,ny = (math.prod(tmask_shape[0:2])), math.prod(tmask_shape[2:])
-        tiles_mask = (tiles_mask > 0)
-        tiles_mask = tiles_mask.reshape((n,ny)).all(axis = 1)
-
-        # Get RGB tiles in single array and filter by the mask
-        tshape = tiles.shape
-        tiles_all = tiles.reshape((math.prod(tiles.shape[0:2]), * tiles.shape[2:]))
-        filtered_tiles = tiles_all[tiles_mask]
-
-        if len(filtered_tiles) == 0:
-            return im
-
-        # Find all invalid tiles and replace with a random valid tile
-        replace_count = (tiles_mask == False).sum()
-        rng = np.random.default_rng(seed = seed)
-        tiles_all[np.logical_not(tiles_mask)] = filtered_tiles[rng.choice(filtered_tiles.shape[0], replace_count),:,:,:]
-
-        # Convert back to an image
-        tiles_all = tiles_all.reshape(tshape)
-        tiles_all = tiles_all.swapaxes(1,2)
-        st = tiles_all.reshape((math.prod(tiles_all.shape[0:2]), math.prod(tiles_all.shape[2:4]), tiles_all.shape[4]))
-        si = Image.fromarray(st, mode='RGBA')
-
-        return si
-
-
-    def mask_edge(self, mask: Image, edge_size: int, edge_blur: int) -> Image:
-        npimg = np.asarray(mask, dtype=np.uint8)
-
-        # Detect any partially transparent regions
-        npgradient = np.uint8(255 * (1.0 - np.floor(np.abs(0.5 - np.float32(npimg) / 255.0) * 2.0)))
-
-        # Detect hard edges
-        npedge = cv.Canny(npimg, threshold1=100, threshold2=200)
-
-        # Combine
-        npmask = npgradient + npedge
-
-        # Expand
-        npmask = cv.dilate(npmask, np.ones((3,3), np.uint8), iterations = int(edge_size / 2))
-
-        new_mask = Image.fromarray(npmask)
-
-        if edge_blur > 0:
-            new_mask = new_mask.filter(ImageFilter.BoxBlur(edge_blur))
-
-        return ImageOps.invert(new_mask)
-
-
-    def seam_paint(self,
-        im: Image.Image,
-        seam_size: int,
-        seam_blur: int,
-        prompt,sampler,steps,cfg_scale,ddim_eta,
-        conditioning,strength,
-        noise,
-        step_callback
-    ) -> Image.Image:
-        hard_mask = self.pil_image.split()[-1].copy()
-        mask = self.mask_edge(hard_mask, seam_size, seam_blur)
-
-        make_image = self.get_make_image(
-            prompt,
-            sampler,
-            steps,
-            cfg_scale,
-            ddim_eta,
-            conditioning,
-            init_image = im.copy().convert('RGBA'),
-            mask_image = mask.convert('RGB'), # Code currently requires an RGB mask
-            strength = strength,
-            mask_blur_radius = 0,
-            seam_size = 0,
-            step_callback = step_callback,
-            inpaint_width = im.width,
-            inpaint_height = im.height
-        )
-
-        seam_noise = self.get_noise(im.width, im.height)
-
-        result = make_image(seam_noise)
-
-        return result
-
-
-    @torch.no_grad()
-    def get_make_image(self,prompt,sampler,steps,cfg_scale,ddim_eta,
-                       conditioning,init_image,mask_image,strength,
-                       mask_blur_radius: int = 8,
-                       # Seam settings - when 0, doesn't fill seam
-                       seam_size: int = 0,
-                       seam_blur: int = 0,
-                       seam_strength: float = 0.7,
-                       seam_steps: int = 10,
-                       tile_size: int = 32,
-                       step_callback=None,
-                       inpaint_replace=False, enable_image_debugging=False,
-                       infill_method = None,
-                       inpaint_width=None,
-                       inpaint_height=None,
-                       **kwargs):
-        """
-        Returns a function returning an image derived from the prompt and
-        the initial image + mask.  Return value depends on the seed at
-        the time you call it.  kwargs are 'init_latent' and 'strength'
-        """
-
-        self.enable_image_debugging = enable_image_debugging
-        self.infill_method = infill_method or infill_methods()[0], # The infill method to use
-        
-        self.inpaint_width = inpaint_width
-        self.inpaint_height = inpaint_height
-
-        if isinstance(init_image, PIL.Image.Image):
-            self.pil_image = init_image.copy()
-
-            # Do infill
-            if infill_method == 'patchmatch' and PatchMatch.patchmatch_available():
-                init_filled = self.infill_patchmatch(self.pil_image.copy())
-            else: # if infill_method == 'tile': # Only two methods right now, so always use 'tile' if not patchmatch
-                init_filled = self.tile_fill_missing(
-                    self.pil_image.copy(),
-                    seed = self.seed,
-                    tile_size = tile_size
-                )
-            init_filled.paste(init_image, (0,0), init_image.split()[-1])
-
-            # Resize if requested for inpainting
-            if inpaint_width and inpaint_height:
-                init_filled = init_filled.resize((inpaint_width, inpaint_height))
-
-            debug_image(init_filled, "init_filled", debug_status=self.enable_image_debugging)
-
-            # Create init tensor
-            init_image = self._image_to_tensor(init_filled.convert('RGB'))
-
-        if isinstance(mask_image, PIL.Image.Image):
-            self.pil_mask = mask_image.copy()
-            debug_image(mask_image, "mask_image BEFORE multiply with pil_image", debug_status=self.enable_image_debugging)
-
-            mask_image = ImageChops.multiply(mask_image, self.pil_image.split()[-1].convert('RGB'))
-            self.pil_mask = mask_image
-
-            # Resize if requested for inpainting
-            if inpaint_width and inpaint_height:
-                mask_image = mask_image.resize((inpaint_width, inpaint_height))
-
-            debug_image(mask_image, "mask_image AFTER multiply with pil_image", debug_status=self.enable_image_debugging)
-            mask_image = mask_image.resize(
-                (
-                    mask_image.width // downsampling,
-                    mask_image.height // downsampling
-                ),
-                resample=Image.Resampling.NEAREST
-            )
-            mask_image = self._image_to_tensor(mask_image,normalize=False)
-
-        self.mask_blur_radius = mask_blur_radius
-
-        # klms samplers not supported yet, so ignore previous sampler
-        if isinstance(sampler,KSampler):
-            print(
-                f">> Using recommended DDIM sampler for inpainting."
-            )
-            sampler = DDIMSampler(self.model, device=self.model.device)
-
-        sampler.make_schedule(
-            ddim_num_steps=steps, ddim_eta=ddim_eta, verbose=False
-        )
-
-        mask_image = mask_image[0][0].unsqueeze(0).repeat(4,1,1).unsqueeze(0)
-        mask_image = repeat(mask_image, '1 ... -> b ...', b=1)
-
-        scope = choose_autocast(self.precision)
-        with scope(self.model.device.type):
-            self.init_latent = self.model.get_first_stage_encoding(
-                self.model.encode_first_stage(init_image)
-            ) # move to latent space
-
-        t_enc   = int(strength * steps)
-        # todo: support cross-attention control
-        uc, c, _ = conditioning
-
-        print(f">> target t_enc is {t_enc} steps")
-
-        @torch.no_grad()
-        def make_image(x_T):
-            # encode (scaled latent)
-            z_enc = sampler.stochastic_encode(
-                self.init_latent,
-                torch.tensor([t_enc - 1]).to(self.model.device),
-                noise=x_T
-            )
-
-            # to replace masked area with latent noise, weighted by inpaint_replace strength
-            if inpaint_replace > 0.0:
-                print(f'>> inpaint will replace what was under the mask with a strength of {inpaint_replace}')
-                l_noise = self.get_noise(kwargs['width'],kwargs['height'])
-                inverted_mask = 1.0-mask_image  # there will be 1s where the mask is
-                masked_region = (1.0-inpaint_replace) * inverted_mask * z_enc + inpaint_replace * inverted_mask * l_noise
-                z_enc   = z_enc * mask_image + masked_region
-
-            if self.free_gpu_mem and self.model.model.device != self.model.device:
-                self.model.model.to(self.model.device)
-
-            # decode it
-            samples = sampler.decode(
-                z_enc,
-                c,
-                t_enc,
-                img_callback                 = step_callback,
-                unconditional_guidance_scale = cfg_scale,
-                unconditional_conditioning = uc,
-                mask                       = mask_image,
-                init_latent                = self.init_latent
-            )
-
-            result = self.sample_to_image(samples)
-
-            # Seam paint if this is our first pass (seam_size set to 0 during seam painting)
-            if seam_size > 0:
-                old_image = self.pil_image or init_image
-                old_mask = self.pil_mask or mask_image
-
-                result = self.seam_paint(
-                    result,
-                    seam_size,
-                    seam_blur,
-                    prompt,
-                    sampler,
-                    seam_steps,
-                    cfg_scale,
-                    ddim_eta,
-                    conditioning,
-                    seam_strength,
-                    x_T,
-                    step_callback)
-
-                # Restore original settings
-                self.get_make_image(prompt,sampler,steps,cfg_scale,ddim_eta,
-                       conditioning,
-                       old_image,
-                       old_mask,
-                       strength,
-                       mask_blur_radius, seam_size, seam_blur, seam_strength,
-                       seam_steps, tile_size, step_callback,
-                       inpaint_replace, enable_image_debugging,
-                       inpaint_width = inpaint_width,
-                       inpaint_height = inpaint_height,
-                       infill_method = infill_method,
-                       **kwargs)
-
-            return result
-
-        return make_image
-
-
-    def sample_to_image(self, samples)->Image.Image:
-        gen_result = super().sample_to_image(samples).convert('RGB')
-        debug_image(gen_result, "gen_result", debug_status=self.enable_image_debugging)
-
-        # Resize if necessary
-        if self.inpaint_width and self.inpaint_height:
-            gen_result = gen_result.resize(self.pil_image.size)
-
-        if self.pil_image is None or self.pil_mask is None:
-            return gen_result
-
-        corrected_result = super().repaste_and_color_correct(gen_result, self.pil_image, self.pil_mask, self.mask_blur_radius)
-        debug_image(corrected_result, "corrected_result", debug_status=self.enable_image_debugging)
-
-        return corrected_result

ldm/invoke/ckpt_generator/omnibus.py

@@ -1,175 +0,0 @@
-"""omnibus module to be used with the runwayml 9-channel custom inpainting model"""
-
-import torch
-import numpy as  np
-from einops import repeat
-from PIL import Image, ImageOps, ImageChops
-from ldm.invoke.devices import choose_autocast
-from ldm.invoke.ckpt_generator.base import downsampling
-from ldm.invoke.ckpt_generator.img2img import CkptImg2Img
-from ldm.invoke.ckpt_generator.txt2img import CkptTxt2Img
-
-class CkptOmnibus(CkptImg2Img,CkptTxt2Img):
-    def __init__(self, model, precision):
-        super().__init__(model, precision)
-        self.pil_mask = None
-        self.pil_image = None
-
-    def get_make_image(
-            self,
-            prompt,
-            sampler,
-            steps,
-            cfg_scale,
-            ddim_eta,
-            conditioning,
-            width,
-            height,
-            init_image = None,
-            mask_image = None,
-            strength = None,
-            step_callback=None,
-            threshold=0.0,
-            perlin=0.0,
-            mask_blur_radius: int = 8,
-            **kwargs):
-        """
-        Returns a function returning an image derived from the prompt and the initial image
-        Return value depends on the seed at the time you call it.
-        """
-        self.perlin = perlin
-        num_samples = 1
-
-        sampler.make_schedule(
-            ddim_num_steps=steps, ddim_eta=ddim_eta, verbose=False
-        )
-
-        if isinstance(init_image, Image.Image):
-            self.pil_image = init_image
-            if init_image.mode != 'RGB':
-                init_image = init_image.convert('RGB')
-            init_image = self._image_to_tensor(init_image)
-
-        if isinstance(mask_image, Image.Image):
-            self.pil_mask = mask_image
-
-            mask_image = ImageChops.multiply(mask_image.convert('L'), self.pil_image.split()[-1])
-            mask_image = self._image_to_tensor(ImageOps.invert(mask_image), normalize=False)
-
-        self.mask_blur_radius = mask_blur_radius
-
-        t_enc = steps
-
-        if init_image is not None and mask_image is not None: # inpainting
-            masked_image = init_image * (1 - mask_image)  # masked image is the image masked by mask - masked regions zero
-
-        elif init_image is not None: # img2img
-            scope = choose_autocast(self.precision)
-
-            with scope(self.model.device.type):
-                self.init_latent = self.model.get_first_stage_encoding(
-                    self.model.encode_first_stage(init_image)
-                ) # move to latent space
-
-            # create a completely black mask  (1s)
-            mask_image = torch.ones(1, 1, init_image.shape[2], init_image.shape[3], device=self.model.device)
-            # and the masked image is just a copy of the original
-            masked_image = init_image
-
-        else: # txt2img
-            init_image = torch.zeros(1, 3, height, width, device=self.model.device)
-            mask_image = torch.ones(1, 1, height, width, device=self.model.device)
-            masked_image = init_image
-
-        self.init_latent = init_image
-        height = init_image.shape[2]
-        width = init_image.shape[3]
-        model = self.model
-
-        def make_image(x_T):
-            with torch.no_grad():
-                scope = choose_autocast(self.precision)
-                with scope(self.model.device.type):
-
-                    batch = self.make_batch_sd(
-                        init_image,
-                        mask_image,
-                        masked_image,
-                        prompt=prompt,
-                        device=model.device,
-                        num_samples=num_samples,
-                    )
-
-                    c = model.cond_stage_model.encode(batch["txt"])
-                    c_cat = list()
-                    for ck in model.concat_keys:
-                        cc = batch[ck].float()
-                        if ck != model.masked_image_key:
-                            bchw = [num_samples, 4, height//8, width//8]
-                            cc = torch.nn.functional.interpolate(cc, size=bchw[-2:])
-                        else:
-                            cc = model.get_first_stage_encoding(model.encode_first_stage(cc))
-                        c_cat.append(cc)
-                    c_cat = torch.cat(c_cat, dim=1)
-
-                    # cond
-                    cond={"c_concat": [c_cat], "c_crossattn": [c]}
-
-                    # uncond cond
-                    uc_cross = model.get_unconditional_conditioning(num_samples, "")
-                    uc_full = {"c_concat": [c_cat], "c_crossattn": [uc_cross]}
-                    shape = [model.channels, height//8, width//8]
-
-                    samples, _ = sampler.sample(
-                        batch_size = 1,
-                        S = steps,
-                        x_T = x_T,
-                        conditioning = cond,
-                        shape = shape,
-                        verbose = False,
-                        unconditional_guidance_scale = cfg_scale,
-                        unconditional_conditioning = uc_full,
-                        eta = 1.0,
-                        img_callback = step_callback,
-                        threshold = threshold,
-                    )
-                    if self.free_gpu_mem:
-                        self.model.model.to("cpu")
-            return self.sample_to_image(samples)
-
-        return make_image
-
-    def make_batch_sd(
-            self,
-            image,
-            mask,
-            masked_image,
-            prompt,
-            device,
-            num_samples=1):
-        batch = {
-                "image": repeat(image.to(device=device), "1 ... -> n ...", n=num_samples),
-                "txt": num_samples * [prompt],
-                "mask": repeat(mask.to(device=device), "1 ... -> n ...", n=num_samples),
-                "masked_image": repeat(masked_image.to(device=device), "1 ... -> n ...", n=num_samples),
-                }
-        return batch
-
-    def get_noise(self, width:int, height:int):
-        if self.init_latent is not None:
-            height = self.init_latent.shape[2]
-            width = self.init_latent.shape[3]
-        return CkptTxt2Img.get_noise(self,width,height)
-
-
-    def sample_to_image(self, samples)->Image.Image:
-        gen_result = super().sample_to_image(samples).convert('RGB')
-
-        if self.pil_image is None or self.pil_mask is None:
-            return gen_result
-        if self.pil_image.size != self.pil_mask.size:
-            return gen_result
-
-        corrected_result = super(CkptImg2Img, self).repaste_and_color_correct(gen_result, self.pil_image, self.pil_mask, self.mask_blur_radius)
-
-        return corrected_result

ldm/invoke/ckpt_generator/txt2img.py

@@ -1,90 +0,0 @@
-'''
-ldm.invoke.ckpt_generator.txt2img inherits from ldm.invoke.ckpt_generator
-'''
-
-import torch
-import numpy as  np
-from ldm.invoke.ckpt_generator.base import CkptGenerator
-from ldm.models.diffusion.shared_invokeai_diffusion import InvokeAIDiffuserComponent
-import gc
-
-
-class CkptTxt2Img(CkptGenerator):
-    def __init__(self, model, precision):
-        super().__init__(model, precision)
-
-    @torch.no_grad()
-    def get_make_image(self,prompt,sampler,steps,cfg_scale,ddim_eta,
-                       conditioning,width,height,step_callback=None,threshold=0.0,perlin=0.0,
-                       attention_maps_callback=None,
-                       **kwargs):
-        """
-        Returns a function returning an image derived from the prompt and the initial image
-        Return value depends on the seed at the time you call it
-        kwargs are 'width' and 'height'
-        """
-        self.perlin = perlin
-        uc, c, extra_conditioning_info   = conditioning
-
-        @torch.no_grad()
-        def make_image(x_T):
-            shape = [
-                self.latent_channels,
-                height // self.downsampling_factor,
-                width  // self.downsampling_factor,
-            ]
-
-            if self.free_gpu_mem and self.model.model.device != self.model.device:
-                self.model.model.to(self.model.device)
-
-            sampler.make_schedule(ddim_num_steps=steps, ddim_eta=ddim_eta, verbose=False)
-
-            samples, _ = sampler.sample(
-                batch_size                   = 1,
-                S                            = steps,
-                x_T                          = x_T,
-                conditioning                 = c,
-                shape                        = shape,
-                verbose                      = False,
-                unconditional_guidance_scale = cfg_scale,
-                unconditional_conditioning   = uc,
-                extra_conditioning_info      = extra_conditioning_info,
-                eta                          = ddim_eta,
-                img_callback                 = step_callback,
-                threshold                    = threshold,
-                attention_maps_callback      = attention_maps_callback,
-            )
-
-            if self.free_gpu_mem:
-                self.model.model.to('cpu')
-                self.model.cond_stage_model.device = 'cpu'
-                self.model.cond_stage_model.to('cpu')
-                gc.collect()
-                torch.cuda.empty_cache()
-
-            return self.sample_to_image(samples)
-
-        return make_image
-
-
-    # returns a tensor filled with random numbers from a normal distribution
-    def get_noise(self,width,height):
-        device         = self.model.device
-        if self.use_mps_noise or device.type == 'mps':
-            x = torch.randn([1,
-                             self.latent_channels,
-                             height // self.downsampling_factor,
-                             width  // self.downsampling_factor],
-                            dtype=self.torch_dtype(),
-                            device='cpu').to(device)
-        else:
-            x = torch.randn([1,
-                             self.latent_channels,
-                             height // self.downsampling_factor,
-                             width  // self.downsampling_factor],
-                            dtype=self.torch_dtype(),
-                            device=device)
-        if self.perlin > 0.0:
-            x = (1-self.perlin)*x + self.perlin*self.get_perlin_noise(width  // self.downsampling_factor, height // self.downsampling_factor)
-        return x
-

ldm/invoke/ckpt_generator/txt2img2img.py

@@ -1,182 +0,0 @@
-'''
-ldm.invoke.ckpt_generator.txt2img inherits from ldm.invoke.ckpt_generator
-'''
-
-import torch
-import numpy as  np
-import math
-import gc
-from ldm.invoke.ckpt_generator.base import CkptGenerator
-from ldm.invoke.ckpt_generator.omnibus import CkptOmnibus
-from ldm.models.diffusion.ddim import DDIMSampler
-from ldm.models.diffusion.shared_invokeai_diffusion import InvokeAIDiffuserComponent
-from PIL import Image
-
-class CkptTxt2Img2Img(CkptGenerator):
-    def __init__(self, model, precision):
-        super().__init__(model, precision)
-        self.init_latent = None    # for get_noise()
-
-    @torch.no_grad()
-    def get_make_image(self,prompt,sampler,steps,cfg_scale,ddim_eta,
-                       conditioning,width,height,strength,step_callback=None,**kwargs):
-        """
-        Returns a function returning an image derived from the prompt and the initial image
-        Return value depends on the seed at the time you call it
-        kwargs are 'width' and 'height'
-        """
-        uc, c, extra_conditioning_info = conditioning
-        scale_dim = min(width, height)
-        scale = 512 / scale_dim
-
-        init_width = math.ceil(scale * width / 64) * 64
-        init_height = math.ceil(scale * height / 64) * 64
-
-        @torch.no_grad()
-        def make_image(x_T):
-
-            shape = [
-                self.latent_channels,
-                init_height // self.downsampling_factor,
-                init_width // self.downsampling_factor,
-            ]
-
-            sampler.make_schedule(
-                    ddim_num_steps=steps, ddim_eta=ddim_eta, verbose=False
-            )
-
-            #x = self.get_noise(init_width, init_height)
-            x = x_T
-
-            if self.free_gpu_mem and self.model.model.device != self.model.device:
-                self.model.model.to(self.model.device)
-
-            samples, _ = sampler.sample(
-                batch_size                   = 1,
-                S                            = steps,
-                x_T                          = x,
-                conditioning                 = c,
-                shape                        = shape,
-                verbose                      = False,
-                unconditional_guidance_scale = cfg_scale,
-                unconditional_conditioning   = uc,
-                eta                          = ddim_eta,
-                img_callback                 = step_callback,
-                extra_conditioning_info      = extra_conditioning_info
-            )
-
-            print(
-                  f"\n>> Interpolating from {init_width}x{init_height} to {width}x{height} using DDIM sampling"
-                 )
-
-            # resizing
-            samples = torch.nn.functional.interpolate(
-                samples,
-                size=(height // self.downsampling_factor, width // self.downsampling_factor),
-                mode="bilinear"
-            )
-
-            t_enc = int(strength * steps)
-            ddim_sampler = DDIMSampler(self.model, device=self.model.device)
-            ddim_sampler.make_schedule(
-                    ddim_num_steps=steps, ddim_eta=ddim_eta, verbose=False
-            )
-
-            z_enc = ddim_sampler.stochastic_encode(
-                samples,
-                torch.tensor([t_enc-1]).to(self.model.device),
-                noise=self.get_noise(width,height,False)
-            )
-
-            # decode it
-            samples = ddim_sampler.decode(
-                z_enc,
-                c,
-                t_enc,
-                img_callback = step_callback,
-                unconditional_guidance_scale=cfg_scale,
-                unconditional_conditioning=uc,
-                extra_conditioning_info=extra_conditioning_info,
-                all_timesteps_count=steps
-            )
-
-            if self.free_gpu_mem:
-                self.model.model.to('cpu')
-                self.model.cond_stage_model.device = 'cpu'
-                self.model.cond_stage_model.to('cpu')
-                gc.collect()
-                torch.cuda.empty_cache()
-
-            return self.sample_to_image(samples)
-
-        # in the case of the inpainting model being loaded, the trick of
-        # providing an interpolated latent doesn't work, so we transiently
-        # create a 512x512 PIL image, upscale it, and run the inpainting
-        # over it in img2img mode. Because the inpaing model is so conservative
-        # it doesn't change the image (much)
-        def inpaint_make_image(x_T):
-            omnibus = CkptOmnibus(self.model,self.precision)
-            result = omnibus.generate(
-                prompt,
-                sampler=sampler,
-                width=init_width,
-                height=init_height,
-                step_callback=step_callback,
-                steps = steps,
-                cfg_scale = cfg_scale,
-                ddim_eta = ddim_eta,
-                conditioning = conditioning,
-                **kwargs
-            )
-            assert result is not None and len(result)>0,'** txt2img failed **'
-            image = result[0][0]
-            interpolated_image = image.resize((width,height),resample=Image.Resampling.LANCZOS)
-            print(kwargs.pop('init_image',None))
-            result = omnibus.generate(
-                prompt,
-                sampler=sampler,
-                init_image=interpolated_image,
-                width=width,
-                height=height,
-                seed=result[0][1],
-                step_callback=step_callback,
-                steps = steps,
-                cfg_scale = cfg_scale,
-                ddim_eta = ddim_eta,
-                conditioning = conditioning,
-                **kwargs
-                )
-            return result[0][0]
-
-        if sampler.uses_inpainting_model():
-            return inpaint_make_image
-        else:
-            return make_image
-
-    # returns a tensor filled with random numbers from a normal distribution
-    def get_noise(self,width,height,scale = True):
-        # print(f"Get noise: {width}x{height}")
-        if scale:
-            trained_square = 512 * 512
-            actual_square = width * height
-            scale = math.sqrt(trained_square / actual_square)
-            scaled_width = math.ceil(scale * width / 64) * 64
-            scaled_height = math.ceil(scale * height / 64) * 64
-        else:
-            scaled_width = width
-            scaled_height = height
-
-        device      = self.model.device
-        if self.use_mps_noise or device.type == 'mps':
-            return torch.randn([1,
-                                self.latent_channels,
-                                scaled_height // self.downsampling_factor,
-                                scaled_width  // self.downsampling_factor],
-                                device='cpu').to(device)
-        else:
-            return torch.randn([1,
-                                self.latent_channels,
-                                scaled_height // self.downsampling_factor,
-                                scaled_width  // self.downsampling_factor],
-                                device=device)
-

ldm/invoke/ckpt_to_diffuser.py

@@ -1037,10 +1037,10 @@ def convert_open_clip_checkpoint(checkpoint):
     return text_model
 
 def replace_checkpoint_vae(checkpoint, vae_path:str):
-    if vae_path.endswith(".safetensors"):
-        vae_ckpt = load_file(vae_path)
-    else:
+    if Path(vae_path).suffix in ['.pt','.ckpt']:
         vae_ckpt = torch.load(vae_path, map_location="cpu")
+    else:
+        vae_ckpt = load_file(vae_path)
     state_dict = vae_ckpt['state_dict'] if "state_dict" in vae_ckpt else vae_ckpt
     for vae_key in state_dict:
         new_key = f'first_stage_model.{vae_key}'
@@ -1264,10 +1264,10 @@ def load_pipeline_from_original_stable_diffusion_ckpt(
                 cache_dir=cache_dir,
             )
             pipe = pipeline_class(
-                vae=vae,
-                text_encoder=text_model,
+                vae=vae.to(precision),
+                text_encoder=text_model.to(precision),
                 tokenizer=tokenizer,
-                unet=unet,
+                unet=unet.to(precision),
                 scheduler=scheduler,
                 safety_checker=None,
                 feature_extractor=None,

ldm/invoke/concepts_lib.py

@@ -24,13 +24,15 @@ class HuggingFaceConceptsLibrary(object):
         self.concepts_loaded = dict()
         self.triggers = dict()            # concept name to trigger phrase
         self.concept_names = dict()       # trigger phrase to concept name
-        self.match_trigger = re.compile('(<[\w\- >]+>)') # trigger is slightly less restrictive than HF concept name
-        self.match_concept = re.compile('<([\w\-]+)>') # HF concept name can only contain A-Za-z0-9_-
+        self.match_trigger = re.compile('(<[a-zA-Z0-9_\- >]+>)') # trigger is slightly less restrictive than HF concept name
+        self.match_concept = re.compile('<([a-zA-Z0-9_\-]+)>') # HF concept name can only contain A-Za-z0-9_-
 
-    def list_concepts(self)->list:
+    def list_concepts(self, minimum_likes: int=0)->list:
         '''
         Return a list of all the concepts by name, without the 'sd-concepts-library' part.
         Also adds local concepts in invokeai/embeddings folder.
+        If minimum_likes is provided, then only concepts that have received at least that
+        many "likes" will be returned.
         '''
         local_concepts_now = self.get_local_concepts(os.path.join(self.root, 'embeddings'))
         local_concepts_to_add = set(local_concepts_now).difference(set(self.local_concepts))
@@ -44,7 +46,7 @@ class HuggingFaceConceptsLibrary(object):
         else:
             try:
                 models = self.hf_api.list_models(filter=ModelFilter(model_name='sd-concepts-library/'))
-                self.concept_list = [a.id.split('/')[1] for a in models]
+                self.concept_list = [a.id.split('/')[1] for a in models if a.likes>=minimum_likes]
                 # when init, add all in dir. when not init, add only concepts added between init and now
                 self.concept_list.extend(list(local_concepts_to_add))
             except Exception as e:
@@ -181,7 +183,7 @@ class HuggingFaceConceptsLibrary(object):
 
         print(f'>> Downloading {repo_id}...',end='')
         try:
-            for file in ('README.md','learned_embeds.bin','token_identifier.txt','type_of_concept.txt'):
+            for file in ('learned_embeds.bin','token_identifier.txt'):
                 url = hf_hub_url(repo_id, file)
                 request.urlretrieve(url, os.path.join(dest,file),reporthook=tally_download_size)
         except ul_error.HTTPError as e:

ldm/invoke/conditioning.py

@@ -12,21 +12,13 @@ from typing import Union, Optional, Any
 from transformers import CLIPTokenizer
 
 from compel import Compel
-from compel.prompt_parser import FlattenedPrompt, Blend, Fragment, CrossAttentionControlSubstitute, PromptParser
+from compel.prompt_parser import FlattenedPrompt, Blend, Fragment, CrossAttentionControlSubstitute, PromptParser, \
+    Conjunction
 from .devices import torch_dtype
+from .generator.diffusers_pipeline import StableDiffusionGeneratorPipeline
 from ..models.diffusion.shared_invokeai_diffusion import InvokeAIDiffuserComponent
 from ldm.invoke.globals import Globals
 
-def get_tokenizer(model) -> CLIPTokenizer:
-    # TODO remove legacy ckpt fallback handling
-    return (getattr(model, 'tokenizer', None) # diffusers
-            or model.cond_stage_model.tokenizer) # ldm
-
-def get_text_encoder(model) -> Any:
-    # TODO remove legacy ckpt fallback handling
-    return (getattr(model, 'text_encoder', None)  # diffusers
-            or UnsqueezingLDMTransformer(model.cond_stage_model.transformer)) # ldm
-
 class UnsqueezingLDMTransformer:
     def __init__(self, ldm_transformer):
         self.ldm_transformer = ldm_transformer
@@ -40,40 +32,61 @@ class UnsqueezingLDMTransformer:
         return insufficiently_unsqueezed_tensor.unsqueeze(0)
 
 
-def get_uc_and_c_and_ec(prompt_string, model, log_tokens=False, skip_normalize_legacy_blend=False):
+def get_uc_and_c_and_ec(prompt_string,
+                        model: StableDiffusionGeneratorPipeline,
+                        log_tokens=False, skip_normalize_legacy_blend=False):
     # lazy-load any deferred textual inversions.
     # this might take a couple of seconds the first time a textual inversion is used.
     model.textual_inversion_manager.create_deferred_token_ids_for_any_trigger_terms(prompt_string)
 
-    tokenizer = get_tokenizer(model)
-    text_encoder = get_text_encoder(model)
-    compel = Compel(tokenizer=tokenizer,
-                    text_encoder=text_encoder,
+    compel = Compel(tokenizer=model.tokenizer,
+                    text_encoder=model.text_encoder,
                     textual_inversion_manager=model.textual_inversion_manager,
                     dtype_for_device_getter=torch_dtype)
 
     # get rid of any newline characters
     prompt_string = prompt_string.replace("\n", " ")
     positive_prompt_string, negative_prompt_string = split_prompt_to_positive_and_negative(prompt_string)
+
     legacy_blend = try_parse_legacy_blend(positive_prompt_string, skip_normalize_legacy_blend)
-    positive_prompt: FlattenedPrompt|Blend
+    positive_conjunction: Conjunction
     if legacy_blend is not None:
-        positive_prompt = legacy_blend
+        positive_conjunction = legacy_blend
     else:
-        positive_prompt = Compel.parse_prompt_string(positive_prompt_string)
-    negative_prompt: FlattenedPrompt|Blend = Compel.parse_prompt_string(negative_prompt_string)
+        positive_conjunction = Compel.parse_prompt_string(positive_prompt_string)
+    positive_prompt = positive_conjunction.prompts[0]
 
+    should_use_lora_manager = True
+    lora_weights = positive_conjunction.lora_weights
+    lora_conditions = None
+    if model.peft_manager:
+        should_use_lora_manager = model.peft_manager.should_use(lora_weights)
+        if not should_use_lora_manager:
+            model.peft_manager.set_loras(lora_weights)
+    if model.lora_manager and should_use_lora_manager:
+        lora_conditions = model.lora_manager.set_loras_conditions(lora_weights)
+
+    negative_conjunction = Compel.parse_prompt_string(negative_prompt_string)
+    negative_prompt: FlattenedPrompt | Blend = negative_conjunction.prompts[0]
+
+    tokens_count = get_max_token_count(model.tokenizer, positive_prompt)
     if log_tokens or getattr(Globals, "log_tokenization", False):
-        log_tokenization(positive_prompt, negative_prompt, tokenizer=tokenizer)
+        log_tokenization(positive_prompt, negative_prompt, tokenizer=model.tokenizer)
 
-    c, options = compel.build_conditioning_tensor_for_prompt_object(positive_prompt)
-    uc, _ = compel.build_conditioning_tensor_for_prompt_object(negative_prompt)
-
-    tokens_count = get_max_token_count(tokenizer, positive_prompt)
+    # some LoRA models also mess with the text encoder, so they must be active while compel builds conditioning tensors
+    lora_conditioning_ec = InvokeAIDiffuserComponent.ExtraConditioningInfo(tokens_count_including_eos_bos=tokens_count,
+                                                                                            lora_conditions=lora_conditions)
+    with InvokeAIDiffuserComponent.custom_attention_context(model.unet,
+                                                            extra_conditioning_info=lora_conditioning_ec,
+                                                            step_count=-1):
+        c, options = compel.build_conditioning_tensor_for_prompt_object(positive_prompt)
+        uc, _ = compel.build_conditioning_tensor_for_prompt_object(negative_prompt)
 
+    # now build the "real" ec
     ec = InvokeAIDiffuserComponent.ExtraConditioningInfo(tokens_count_including_eos_bos=tokens_count,
                                                          cross_attention_control_args=options.get(
-                                                             'cross_attention_control', None))
+                                                             'cross_attention_control', None),
+                                                         lora_conditions=lora_conditions)
     return uc, c, ec
 
 
@@ -81,12 +94,14 @@ def get_prompt_structure(prompt_string, skip_normalize_legacy_blend: bool = Fals
     Union[FlattenedPrompt, Blend], FlattenedPrompt):
     positive_prompt_string, negative_prompt_string = split_prompt_to_positive_and_negative(prompt_string)
     legacy_blend = try_parse_legacy_blend(positive_prompt_string, skip_normalize_legacy_blend)
-    positive_prompt: FlattenedPrompt|Blend
+    positive_conjunction: Conjunction
     if legacy_blend is not None:
-        positive_prompt = legacy_blend
+        positive_conjunction = legacy_blend
     else:
-        positive_prompt = Compel.parse_prompt_string(positive_prompt_string)
-    negative_prompt: FlattenedPrompt|Blend = Compel.parse_prompt_string(negative_prompt_string)
+        positive_conjunction = Compel.parse_prompt_string(positive_prompt_string)
+    positive_prompt = positive_conjunction.prompts[0]
+    negative_conjunction = Compel.parse_prompt_string(negative_prompt_string)
+    negative_prompt: FlattenedPrompt|Blend = negative_conjunction.prompts[0]
 
     return positive_prompt, negative_prompt
 
@@ -203,18 +218,26 @@ def log_tokenization_for_text(text, tokenizer, display_label=None):
         print(f'{discarded}\x1b[0m')
 
 
-def try_parse_legacy_blend(text: str, skip_normalize: bool=False) -> Optional[Blend]:
+def try_parse_legacy_blend(text: str, skip_normalize: bool=False) -> Optional[Conjunction]:
     weighted_subprompts = split_weighted_subprompts(text, skip_normalize=skip_normalize)
     if len(weighted_subprompts) <= 1:
         return None
     strings = [x[0] for x in weighted_subprompts]
-    weights = [x[1] for x in weighted_subprompts]
 
     pp = PromptParser()
     parsed_conjunctions = [pp.parse_conjunction(x) for x in strings]
-    flattened_prompts = [x.prompts[0] for x in parsed_conjunctions]
+    flattened_prompts = []
+    weights = []
+    loras = []
+    for i, x in enumerate(parsed_conjunctions):
+        if len(x.prompts)>0:
+            flattened_prompts.append(x.prompts[0])
+            weights.append(weighted_subprompts[i][1])
+        if len(x.lora_weights)>0:
+            loras.extend(x.lora_weights)
 
-    return Blend(prompts=flattened_prompts, weights=weights, normalize_weights=not skip_normalize)
+    return Conjunction([Blend(prompts=flattened_prompts, weights=weights, normalize_weights=not skip_normalize)],
+                       lora_weights = loras)
 
 
 def split_weighted_subprompts(text, skip_normalize=False)->list:

ldm/invoke/config/invokeai_configure.py

@@ -655,6 +655,7 @@ def initialize_rootdir(root: str, yes_to_all: bool = False):
         "models",
         "configs",
         "embeddings",
+        "loras",
         "text-inversion-output",
         "text-inversion-training-data",
     ):

ldm/invoke/config/invokeai_update.py

@@ -4,18 +4,19 @@ pip install <path_to_git_source>.
 '''
 import os
 import platform
+import psutil
 import requests
 from rich import box, print
-from rich.console import Console, Group, group
+from rich.console import Console, group
 from rich.panel import Panel
 from rich.prompt import Prompt
 from rich.style import Style
-from rich.syntax import Syntax
-from rich.text import Text
 
 from ldm.invoke import __version__
 
 INVOKE_AI_SRC="https://github.com/invoke-ai/InvokeAI/archive"
+INVOKE_AI_TAG="https://github.com/invoke-ai/InvokeAI/archive/refs/tags"
+INVOKE_AI_BRANCH="https://github.com/invoke-ai/InvokeAI/archive/refs/heads"
 INVOKE_AI_REL="https://api.github.com/repos/invoke-ai/InvokeAI/releases"
 
 OS = platform.uname().system
@@ -30,6 +31,19 @@ else:
 def get_versions()->dict:
     return requests.get(url=INVOKE_AI_REL).json()
 
+def invokeai_is_running()->bool:
+    for p in psutil.process_iter():
+        try:
+            cmdline = p.cmdline()
+            matches = [x for x in cmdline if x.endswith(('invokeai','invokeai.exe'))]
+            if matches:
+                print(f':exclamation: [bold red]An InvokeAI instance appears to be running as process {p.pid}[/red bold]')
+                return True
+        except psutil.AccessDenied:
+            continue
+    return False
+        
+    
 def welcome(versions: dict):
     
     @group()
@@ -41,7 +55,8 @@ def welcome(versions: dict):
         yield '[bold yellow]Options:'
         yield f'''[1] Update to the latest official release ([italic]{versions[0]['tag_name']}[/italic])
 [2] Update to the bleeding-edge development version ([italic]main[/italic])
-[3] Manually enter the tag or branch name you wish to update'''
+[3] Manually enter the [bold]tag name[/bold] for the version you wish to update to
+[4] Manually enter the [bold]branch name[/bold] for the version you wish to update to'''        
 
     console.rule()
     print(
@@ -59,20 +74,33 @@ def welcome(versions: dict):
 
 def main():
     versions = get_versions()
+    if invokeai_is_running():
+        print(f':exclamation: [bold red]Please terminate all running instances of InvokeAI before updating.[/red bold]')
+        return
+
     welcome(versions)
 
     tag = None
-    choice = Prompt.ask('Choice:',choices=['1','2','3'],default='1')
+    branch = None
+    release = None
+    choice = Prompt.ask('Choice:',choices=['1','2','3','4'],default='1')
     
     if choice=='1':
-        tag = versions[0]['tag_name']
+        release = versions[0]['tag_name']
     elif choice=='2':
-        tag = 'main'
+        release = 'main'
     elif choice=='3':
-        tag = Prompt.ask('Enter an InvokeAI tag or branch name')
+        tag = Prompt.ask('Enter an InvokeAI tag name')
+    elif choice=='4':
+        branch = Prompt.ask('Enter an InvokeAI branch name')
 
-    print(f':crossed_fingers: Upgrading to [yellow]{tag}[/yellow]')
-    cmd = f'pip install {INVOKE_AI_SRC}/{tag}.zip --use-pep517 --upgrade'
+    print(f':crossed_fingers: Upgrading to [yellow]{tag if tag else release}[/yellow]')
+    if release:
+        cmd = f'pip install {INVOKE_AI_SRC}/{release}.zip --use-pep517 --upgrade'
+    elif tag:
+        cmd = f'pip install {INVOKE_AI_TAG}/{tag}.zip --use-pep517 --upgrade'
+    else:
+        cmd = f'pip install {INVOKE_AI_BRANCH}/{branch}.zip --use-pep517 --upgrade'
     print('')
     print('')
     if os.system(cmd)==0:

ldm/invoke/config/model_install.py

@@ -196,16 +196,6 @@ class addModelsForm(npyscreen.FormMultiPage):
             scroll_exit=True,
         )
         self.nextrely += 1
-        self.convert_models = self.add_widget_intelligent(
-            npyscreen.TitleSelectOne,
-            name="== CONVERT IMPORTED MODELS INTO DIFFUSERS==",
-            values=["Keep original format", "Convert to diffusers"],
-            value=0,
-            begin_entry_at=4,
-            max_height=4,
-            hidden=True,  # will appear when imported models box is edited
-            scroll_exit=True,
-        )
         self.cancel = self.add_widget_intelligent(
             npyscreen.ButtonPress,
             name="CANCEL",
@@ -240,8 +230,6 @@ class addModelsForm(npyscreen.FormMultiPage):
             self.show_directory_fields.addVisibleWhenSelected(i)
 
         self.show_directory_fields.when_value_edited = self._clear_scan_directory
-        self.import_model_paths.when_value_edited = self._show_hide_convert
-        self.autoload_directory.when_value_edited = self._show_hide_convert
 
     def resize(self):
         super().resize()
@@ -252,13 +240,6 @@ class addModelsForm(npyscreen.FormMultiPage):
         if not self.show_directory_fields.value:
             self.autoload_directory.value = ""
 
-    def _show_hide_convert(self):
-        model_paths = self.import_model_paths.value or ""
-        autoload_directory = self.autoload_directory.value or ""
-        self.convert_models.hidden = (
-            len(model_paths) == 0 and len(autoload_directory) == 0
-        )
-
     def _get_starter_model_labels(self) -> List[str]:
         window_width, window_height = get_terminal_size()
         label_width = 25
@@ -318,7 +299,6 @@ class addModelsForm(npyscreen.FormMultiPage):
         .scan_directory: Path to a directory of models to scan and import
         .autoscan_on_startup:  True if invokeai should scan and import at startup time
         .import_model_paths:   list of URLs, repo_ids and file paths to import
-        .convert_to_diffusers: if True, convert legacy checkpoints into diffusers
         """
         # we're using a global here rather than storing the result in the parentapp
         # due to some bug in npyscreen that is causing attributes to be lost
@@ -354,7 +334,6 @@ class addModelsForm(npyscreen.FormMultiPage):
 
         # URLs and the like
         selections.import_model_paths = self.import_model_paths.value.split()
-        selections.convert_to_diffusers = self.convert_models.value[0] == 1
 
 
 class AddModelApplication(npyscreen.NPSAppManaged):
@@ -367,7 +346,6 @@ class AddModelApplication(npyscreen.NPSAppManaged):
             scan_directory=None,
             autoscan_on_startup=None,
             import_model_paths=None,
-            convert_to_diffusers=None,
         )
 
     def onStart(self):
@@ -387,7 +365,6 @@ def process_and_execute(opt: Namespace, selections: Namespace):
     directory_to_scan = selections.scan_directory
     scan_at_startup = selections.autoscan_on_startup
     potential_models_to_install = selections.import_model_paths
-    convert_to_diffusers = selections.convert_to_diffusers
 
     install_requested_models(
         install_initial_models=models_to_install,
@@ -395,7 +372,6 @@ def process_and_execute(opt: Namespace, selections: Namespace):
         scan_directory=Path(directory_to_scan) if directory_to_scan else None,
         external_models=potential_models_to_install,
         scan_at_startup=scan_at_startup,
-        convert_to_diffusers=convert_to_diffusers,
         precision="float32"
         if opt.full_precision
         else choose_precision(torch.device(choose_torch_device())),

ldm/invoke/config/model_install_backend.py

@@ -11,6 +11,7 @@ from tempfile import TemporaryFile
 
 import requests
 from diffusers import AutoencoderKL
+from diffusers import logging as dlogging
 from huggingface_hub import hf_hub_url
 from omegaconf import OmegaConf
 from omegaconf.dictconfig import DictConfig
@@ -29,7 +30,13 @@ Model_dir = "models"
 Weights_dir = "ldm/stable-diffusion-v1/"
 
 # the initial "configs" dir is now bundled in the `invokeai.configs` package
-Dataset_path = Path(configs.__path__[0]) / "INITIAL_MODELS.yaml"
+Dataset_path = None
+for path in configs.__path__:
+    file =Path(path, "INITIAL_MODELS.yaml")
+    if file.exists():
+        Dataset_path = file
+        break
+assert Dataset_path,f"Could not find the file INITIAL_MODELS.yaml in {configs.__path__}"
 
 # initial models omegaconf
 Datasets = None
@@ -62,7 +69,6 @@ def install_requested_models(
         scan_directory: Path = None,
         external_models: List[str] = None,
         scan_at_startup: bool = False,
-        convert_to_diffusers: bool = False,
         precision: str = "float16",
         purge_deleted: bool = False,
         config_file_path: Path = None,
@@ -105,20 +111,20 @@ def install_requested_models(
     if len(external_models)>0:
         print("== INSTALLING EXTERNAL MODELS ==")
         for path_url_or_repo in external_models:
+            print(f'DEBUG: path_url_or_repo = {path_url_or_repo}')
             try:
                 model_manager.heuristic_import(
                     path_url_or_repo,
-                    convert=convert_to_diffusers,
                     config_file_callback=_pick_configuration_file,
                     commit_to_conf=config_file_path
                 )
             except KeyboardInterrupt:
                 sys.exit(-1)
-            except Exception:
-                pass
+            except Exception as e:
+                print(f'An exception has occurred: {str(e)}')
 
     if scan_at_startup and scan_directory.is_dir():
-        argument = '--autoconvert' if convert_to_diffusers else '--autoimport'
+        argument = '--autoconvert'
         initfile = Path(Globals.root, Globals.initfile)
         replacement = Path(Globals.root, f'{Globals.initfile}.new')
         directory = str(scan_directory).replace('\\','/')
@@ -290,13 +296,21 @@ def _download_diffusion_weights(
     mconfig: DictConfig, access_token: str, precision: str = "float32"
 ):
     repo_id = mconfig["repo_id"]
+    revision = mconfig.get('revision',None)
     model_class = (
         StableDiffusionGeneratorPipeline
         if mconfig.get("format", None) == "diffusers"
         else AutoencoderKL
     )
-    extra_arg_list = [{"revision": "fp16"}, {}] if precision == "float16" else [{}]
+    extra_arg_list = [{"revision": revision}] if revision \
+        else [{"revision": "fp16"}, {}] if precision == "float16" \
+             else [{}]
     path = None
+
+    # quench safety checker warnings
+    verbosity = dlogging.get_verbosity()
+    dlogging.set_verbosity_error()
+    
     for extra_args in extra_arg_list:
         try:
             path = download_from_hf(
@@ -312,6 +326,7 @@ def _download_diffusion_weights(
                 print(f"An unexpected error occurred while downloading the model: {e})")
         if path:
             break
+    dlogging.set_verbosity(verbosity)
     return path
 
 
@@ -442,6 +457,8 @@ def new_config_file_contents(
         stanza["description"] = mod["description"]
         stanza["repo_id"] = mod["repo_id"]
         stanza["format"] = mod["format"]
+        if "revision" in mod:
+            stanza["revision"] = mod["revision"]
         # diffusers don't need width and height (probably .ckpt doesn't either)
         # so we no longer require these in INITIAL_MODELS.yaml
         if "width" in mod:
@@ -466,10 +483,9 @@ def new_config_file_contents(
 
         conf[model] = stanza
 
-    # if no default model was chosen, then we select the first
-    # one in the list
+    # if no default model was chosen, then we select the first one in the list
     if not default_selected:
-        conf[list(successfully_downloaded.keys())[0]]["default"] = True
+        conf[list(conf.keys())[0]]["default"] = True
 
     return OmegaConf.to_yaml(conf)
 

ldm/invoke/dynamic_prompts.py

@@ -99,8 +99,9 @@ def expand_prompts(
             sequence = 0
             for command in commands:
                 sequence += 1
-                parent_conn.send(
-                    command + f' --fnformat="dp.{sequence:04}.{{prompt}}.png"'
+                format = _get_fn_format(outdir, sequence)
+                parent_conn.send_bytes(
+                    (command + f' --fnformat="{format}"').encode('utf-8')
                 )
             parent_conn.close()
         else:
@@ -110,7 +111,20 @@ def expand_prompts(
         for p in children:
             p.terminate()
 
-
+def _get_fn_format(directory:str, sequence:int)->str:
+    """
+    Get a filename that doesn't exceed filename length restrictions
+    on the current platform.
+    """
+    try:
+        max_length = os.pathconf(directory,'PC_NAME_MAX')
+    except:
+        max_length = 255
+    prefix = f'dp.{sequence:04}.'
+    suffix = '.png'
+    max_length -= len(prefix)+len(suffix)
+    return f'{prefix}{{prompt:0.{max_length}}}{suffix}'
+            
 class MessageToStdin(object):
     def __init__(self, connection: Connection):
         self.connection = connection
@@ -119,7 +133,7 @@ class MessageToStdin(object):
     def readline(self) -> str:
         try:
             if len(self.linebuffer) == 0:
-                message = self.connection.recv()
+                message = self.connection.recv_bytes().decode('utf-8')
                 self.linebuffer = message.split("\n")
             result = self.linebuffer.pop(0)
             return result

ldm/invoke/generator/diffusers_pipeline.py

@@ -29,6 +29,8 @@ from typing_extensions import ParamSpec
 from ldm.invoke.globals import Globals
 from ldm.models.diffusion.shared_invokeai_diffusion import InvokeAIDiffuserComponent, PostprocessingSettings
 from ldm.modules.textual_inversion_manager import TextualInversionManager
+from ldm.modules.lora_manager import LoraManager
+from ldm.modules.peft_manager import PeftManager
 from ..devices import normalize_device, CPU_DEVICE
 from ..offloading import LazilyLoadedModelGroup, FullyLoadedModelGroup, ModelGroup
 from ...models.diffusion.cross_attention_map_saving import AttentionMapSaver
@@ -289,11 +291,14 @@ class StableDiffusionGeneratorPipeline(StableDiffusionPipeline):
             safety_checker=safety_checker,
             feature_extractor=feature_extractor,
         )
+        self.lora_manager = LoraManager(self)
+        self.peft_manager = PeftManager()
         self.invokeai_diffuser = InvokeAIDiffuserComponent(self.unet, self._unet_forward, is_running_diffusers=True)
         use_full_precision = (precision == 'float32' or precision == 'autocast')
         self.textual_inversion_manager = TextualInversionManager(tokenizer=self.tokenizer,
                                                                  text_encoder=self.text_encoder,
                                                                  full_precision=use_full_precision)
+
         # InvokeAI's interface for text embeddings and whatnot
         self.embeddings_provider = EmbeddingsProvider(
             tokenizer=self.tokenizer,
@@ -462,8 +467,9 @@ class StableDiffusionGeneratorPipeline(StableDiffusionPipeline):
         if additional_guidance is None:
             additional_guidance = []
         extra_conditioning_info = conditioning_data.extra
-        with self.invokeai_diffuser.custom_attention_context(extra_conditioning_info=extra_conditioning_info,
-                                                             step_count=len(self.scheduler.timesteps)
+        with InvokeAIDiffuserComponent.custom_attention_context(self.invokeai_diffuser.model,
+                                                                extra_conditioning_info=extra_conditioning_info,
+                                                                step_count=len(self.scheduler.timesteps)
                                                              ):
 
             yield PipelineIntermediateState(run_id=run_id, step=-1, timestep=self.scheduler.num_train_timesteps,

ldm/invoke/generator/inpaint.py

@@ -255,8 +255,8 @@ class Inpaint(Img2Img):
         pipeline.scheduler = sampler
 
         # todo: support cross-attention control
-        uc, c, _ = conditioning
-        conditioning_data = (ConditioningData(uc, c, cfg_scale)
+        uc, c, extra_conditioning_info = conditioning
+        conditioning_data = (ConditioningData(uc, c, cfg_scale, extra_conditioning_info)
                              .add_scheduler_args_if_applicable(pipeline.scheduler, eta=ddim_eta))
 
 

ldm/invoke/globals.py

@@ -23,6 +23,7 @@ Globals = Namespace()
 Globals.initfile = 'invokeai.init'
 Globals.models_file = 'models.yaml'
 Globals.models_dir = 'models'
+Globals.lora_models_dir = 'loras'
 Globals.config_dir = 'configs'
 Globals.autoscan_dir = 'weights'
 Globals.converted_ckpts_dir = 'converted_ckpts'
@@ -61,7 +62,7 @@ Globals.sequential_guidance = False
 Globals.full_precision = False
 
 # whether we should convert ckpt files into diffusers models on the fly
-Globals.ckpt_convert = False
+Globals.ckpt_convert = True
 
 # logging tokenization everywhere
 Globals.log_tokenization = False
@@ -75,6 +76,11 @@ def global_config_dir()->Path:
 def global_models_dir()->Path:
     return Path(Globals.root, Globals.models_dir)
 
+def global_lora_models_dir()->Path:
+    return Path(Globals.lora_models_dir) \
+        if Path(Globals.lora_models_dir).is_absolute() \
+           else Path(Globals.root, Globals.lora_models_dir)
+
 def global_autoscan_dir()->Path:
     return Path(Globals.root, Globals.autoscan_dir)
 

ldm/invoke/invokeai_metadata.py

@@ -0,0 +1,29 @@
+#!/usr/bin/env python
+
+import sys
+import json
+from ldm.invoke.pngwriter import retrieve_metadata
+
+def main():
+    if len(sys.argv) < 2:
+        print("Usage: file2prompt.py <file1.png> <file2.png> <file3.png>...")
+        print("This script opens up the indicated invoke.py-generated PNG file(s) and prints out their metadata.")
+        exit(-1)
+
+    filenames = sys.argv[1:]
+    for f in filenames:
+        try:
+            metadata = retrieve_metadata(f)
+            print(f'{f}:\n',json.dumps(metadata['sd-metadata'], indent=4))
+        except FileNotFoundError:
+            sys.stderr.write(f'{f} not found\n')
+            continue
+        except PermissionError:
+            sys.stderr.write(f'{f} could not be opened due to inadequate permissions\n')
+            continue
+
+if __name__ == "__main__":
+    try:
+        main()
+    except KeyboardInterrupt:
+        pass

ldm/invoke/model_manager.py

@@ -19,7 +19,7 @@ import warnings
 from enum import Enum
 from pathlib import Path
 from shutil import move, rmtree
-from typing import Any, Callable, Optional, Union
+from typing import Any, Callable, Optional, Union, List
 
 import safetensors
 import safetensors.torch
@@ -46,12 +46,7 @@ class SDLegacyType(Enum):
     V2_v = 5
     UNKNOWN = 99
 
-
 DEFAULT_MAX_MODELS = 2
-VAE_TO_REPO_ID = {  # hack, see note in convert_and_import()
-    "vae-ft-mse-840000-ema-pruned": "stabilityai/sd-vae-ft-mse",
-}
-
 
 class ModelManager(object):
     def __init__(
@@ -108,11 +103,7 @@ class ModelManager(object):
             requested_model = self.models[model_name]["model"]
             print(f">> Retrieving model {model_name} from system RAM cache")
             self.models[model_name]["model"] = self._model_from_cpu(requested_model)
-            width = self.models[model_name]["width"]
-            height = self.models[model_name]["height"]
-            hash = self.models[model_name]["hash"]
-
-        else:  # we're about to load a new model, so potentially offload the least recently used one
+        else:
             requested_model, width, height, hash = self._load_model(model_name)
             self.models[model_name] = {
                 "model": requested_model,
@@ -123,13 +114,8 @@ class ModelManager(object):
 
         self.current_model = model_name
         self._push_newest_model(model_name)
-        return {
-            "model": requested_model,
-            "width": width,
-            "height": height,
-            "hash": hash,
-        }
-
+        return self.models[model_name]
+    
     def default_model(self) -> str | None:
         """
         Returns the name of the default model, or None
@@ -166,19 +152,6 @@ class ModelManager(object):
         """
         return list(self.config.keys())
 
-    def is_legacy(self, model_name: str) -> bool:
-        """
-        Return true if this is a legacy (.ckpt) model
-        """
-        # if we are converting legacy files automatically, then
-        # there are no legacy ckpts!
-        if Globals.ckpt_convert:
-            return False
-        info = self.model_info(model_name)
-        if "weights" in info and info["weights"].endswith((".ckpt", ".safetensors")):
-            return True
-        return False
-
     def list_models(self) -> dict:
         """
         Return a dict of models in the format:
@@ -379,105 +352,49 @@ class ModelManager(object):
         if not os.path.isabs(weights):
             weights = os.path.normpath(os.path.join(Globals.root, weights))
 
-        # check whether this is a v2 file and force conversion
-        convert = Globals.ckpt_convert or self.is_v2_config(config)
+        if matching_config := self._scan_for_matching_file(Path(weights),suffixes=['.yaml']):
+            print(f'   | Using external config file {matching_config}')
+            config = matching_config
 
-        # if converting automatically to diffusers, then we do the conversion and return
-        # a diffusers pipeline
-        if convert:
-            print(
-                f">> Converting legacy checkpoint {model_name} into a diffusers model..."
-            )
-            from ldm.invoke.ckpt_to_diffuser import load_pipeline_from_original_stable_diffusion_ckpt
-
-            self.offload_model(self.current_model)
-            if vae_config := self._choose_diffusers_vae(model_name):
-                vae = self._load_vae(vae_config)
-            if self._has_cuda():
-                torch.cuda.empty_cache()
-            pipeline = load_pipeline_from_original_stable_diffusion_ckpt(
-                checkpoint_path=weights,
-                original_config_file=config,
-                vae=vae,
-                return_generator_pipeline=True,
-                precision=torch.float16
-                if self.precision == "float16"
-                else torch.float32,
-            )
-            if self.sequential_offload:
-                pipeline.enable_offload_submodels(self.device)
-            else:
-                pipeline.to(self.device)
-
-            return (
-                pipeline,
-                width,
-                height,
-                "NOHASH",
-            )
-
-        # for usage statistics
-        if self._has_cuda():
-            torch.cuda.reset_peak_memory_stats()
-            torch.cuda.empty_cache()
-
-        # this does the work
-        if not os.path.isabs(config):
-            config = os.path.join(Globals.root, config)
-        omega_config = OmegaConf.load(config)
-        with open(weights, "rb") as f:
-            weight_bytes = f.read()
-        model_hash = self._cached_sha256(weights, weight_bytes)
-        sd = None
-
-        if weights.endswith(".ckpt"):
-            self.scan_model(model_name, weights)
-            sd = torch.load(io.BytesIO(weight_bytes), map_location="cpu")
-        else:
-            sd = safetensors.torch.load(weight_bytes)
-
-        del weight_bytes
-        # merged models from auto11 merge board are flat for some reason
-        if "state_dict" in sd:
-            sd = sd["state_dict"]
-
-        print("   | Forcing garbage collection prior to loading new model")
-        gc.collect()
-        model = instantiate_from_config(omega_config.model)
-        model.load_state_dict(sd, strict=False)
-
-        if self.precision == "float16":
-            print("   | Using faster float16 precision")
-            model = model.to(torch.float16)
-        else:
-            print("   | Using more accurate float32 precision")
-
-        # look and load a matching vae file. Code borrowed from AUTOMATIC1111 modules/sd_models.py
+        # get the path to the custom vae, if any
+        vae_path = None
+        # first we use whatever is in the config file
         if vae:
-            if not os.path.isabs(vae):
-                vae = os.path.normpath(os.path.join(Globals.root, vae))
-            if os.path.exists(vae):
-                print(f"   | Loading VAE weights from: {vae}")
-                if vae.endswith((".ckpt", ".pt")):
-                    self.scan_model(vae, vae)
-                    vae_ckpt = torch.load(vae, map_location="cpu")
-                else:
-                    vae_ckpt = safetensors.torch.load_file(vae)
-                vae_dict = {k: v for k, v in vae_ckpt.items() if k[0:4] != "loss"}
-                model.first_stage_model.load_state_dict(vae_dict, strict=False)
-            else:
-                print(f"   | VAE file {vae} not found. Skipping.")
+            path = Path(vae if os.path.isabs(vae) else os.path.normpath(os.path.join(Globals.root, vae)))
+            if path.exists():
+                vae_path = path
+        # then we look for a file with the same basename
+        vae_path = vae_path or self._scan_for_matching_file(Path(weights))
+            
+        # Do the conversion and return a diffusers pipeline
+        print(
+            f">> Converting legacy checkpoint {model_name} into a diffusers model..."
+        )
+        from ldm.invoke.ckpt_to_diffuser import load_pipeline_from_original_stable_diffusion_ckpt
 
-        model.to(self.device)
-        # model.to doesn't change the cond_stage_model.device used to move the tokenizer output, so set it here
-        model.cond_stage_model.device = self.device
+        if self._has_cuda():
+            torch.cuda.empty_cache()
+        pipeline = load_pipeline_from_original_stable_diffusion_ckpt(
+            checkpoint_path=weights,
+            original_config_file=config,
+            vae_path=vae_path,
+            return_generator_pipeline=True,
+            precision=torch.float16
+            if self.precision == "float16"
+            else torch.float32,
+        )
+        if self.sequential_offload:
+            pipeline.enable_offload_submodels(self.device)
+        else:
+            pipeline.to(self.device)
 
-        model.eval()
+        return (
+            pipeline,
+            width,
+            height,
+            "NOHASH",
+        )
 
-        for module in model.modules():
-            if isinstance(module, (torch.nn.Conv2d, torch.nn.ConvTranspose2d)):
-                module._orig_padding_mode = module.padding_mode
-        return model, width, height, model_hash
 
     def _load_diffusers_model(self, mconfig):
         name_or_path = self.model_name_or_path(mconfig)
@@ -500,9 +417,9 @@ class ModelManager(object):
             pipeline_args.update(cache_dir=global_cache_dir("hub"))
         if using_fp16:
             pipeline_args.update(torch_dtype=torch.float16)
-            fp_args_list = [{"revision": "fp16"}, {}]
-        else:
-            fp_args_list = [{}]
+        revision = mconfig.get('revision') or ('fp16' if using_fp16 else None)
+        fp_args_list = [{"revision": revision}] if revision else []
+        fp_args_list.append({})
 
         verbosity = dlogging.get_verbosity()
         dlogging.set_verbosity_error()
@@ -544,6 +461,8 @@ class ModelManager(object):
         return pipeline, width, height, model_hash
 
     def is_v2_config(self, config: Path) -> bool:
+        if not os.path.isabs(config):
+            config = os.path.join(Globals.root, config)
         try:
             mconfig = OmegaConf.load(config)
             return (
@@ -758,7 +677,6 @@ class ModelManager(object):
     def heuristic_import(
         self,
         path_url_or_repo: str,
-        convert: bool = False,
         model_name: str = None,
         description: str = None,
         model_config_file: Path = None,
@@ -780,9 +698,6 @@ class ModelManager(object):
         The model_name and/or description can be provided. If not, they will
         be generated automatically.
 
-        If convert is true, legacy models will be converted to diffusers
-        before importing.
-
         If commit_to_conf is provided, the newly loaded model will be written
         to the `models.yaml` file at the indicated path. Otherwise, the changes
         will only remain in memory.
@@ -818,7 +733,6 @@ class ModelManager(object):
             print(f"  | {thing} appears to be a diffusers file on disk")
             model_name = self.import_diffuser_model(
                 thing,
-                vae=dict(repo_id="stabilityai/sd-vae-ft-mse"),
                 model_name=model_name,
                 description=description,
                 commit_to_conf=commit_to_conf,
@@ -839,7 +753,6 @@ class ModelManager(object):
                 ):
                     if model_name := self.heuristic_import(
                         str(m),
-                        convert,
                         commit_to_conf=commit_to_conf,
                         config_file_callback=config_file_callback,
                     ):
@@ -906,11 +819,11 @@ class ModelManager(object):
                     )
                 elif model_type == SDLegacyType.V2:
                     print(
-                        f"** {thing} is a V2 checkpoint file, but its parameterization cannot be determined. Please provide configuration file path."
+                        f"** {thing} is a V2 checkpoint file, but its parameterization cannot be determined. Please provide the configuration file type or path."
                     )
                 else:
                     print(
-                        f"** {thing} is a legacy checkpoint file but not a known Stable Diffusion model. Please provide configuration file path."
+                        f"** {thing} is a legacy checkpoint file but not a known Stable Diffusion model. Please provide the configuration file type or path."
                     )
 
             if not model_config_file and config_file_callback:
@@ -918,52 +831,28 @@ class ModelManager(object):
             if not model_config_file:
                 return
 
-        if self.is_v2_config(model_config_file):
-            convert = True
-            print("   | This SD-v2 model will be converted to diffusers format for use")
-
-        # look for a custom vae
-        vae_path = None
-        for suffix in ["pt", "ckpt", "safetensors"]:
-            if (model_path.with_suffix(f".vae.{suffix}")).exists():
-                vae_path = model_path.with_suffix(f".vae.{suffix}")
-                print(f"   | Using VAE file {vae_path.name}")
+        if (vae_path := self._scan_for_matching_file(model_path)):
+            print(f"   | Using VAE file {vae_path.name}")
+        
+        diffuser_path = Path(
+            Globals.root, "models", Globals.converted_ckpts_dir, model_path.stem
+        )
         vae = None if vae_path else dict(repo_id="stabilityai/sd-vae-ft-mse")
-
-        if convert:
-            diffuser_path = Path(
-                Globals.root, "models", Globals.converted_ckpts_dir, model_path.stem
-            )
-            model_name = self.convert_and_import(
-                model_path,
-                diffusers_path=diffuser_path,
-                vae=vae,
-                vae_path=vae_path,
-                model_name=model_name,
-                model_description=description,
-                original_config_file=model_config_file,
-                commit_to_conf=commit_to_conf,
-                scan_needed=False,
-            )
-            # in the event that this file was downloaded automatically prior to conversion
-            # we do not keep the original .ckpt/.safetensors around
-            if is_temporary:
-                model_path.unlink(missing_ok=True)
-        else:
-            model_name = self.import_ckpt_model(
-                model_path,
-                config=model_config_file,
-                model_name=model_name,
-                model_description=description,
-                vae=str(
-                    vae_path
-                    or Path(
-                        Globals.root,
-                        "models/ldm/stable-diffusion-v1/vae-ft-mse-840000-ema-pruned.ckpt",
-                    )
-                ),
-                commit_to_conf=commit_to_conf,
-            )
+        model_name = self.convert_and_import(
+            model_path,
+            diffusers_path=diffuser_path,
+            vae=vae,
+            vae_path=vae_path,
+            model_name=model_name,
+            model_description=description,
+            original_config_file=model_config_file,
+            commit_to_conf=commit_to_conf,
+            scan_needed=False,
+        )
+        # in the event that this file was downloaded automatically prior to conversion
+        # we do not keep the original .ckpt/.safetensors around
+        if is_temporary:
+            model_path.unlink(missing_ok=True)
         if commit_to_conf:
             self.commit(commit_to_conf)
         return model_name
@@ -1006,14 +895,17 @@ class ModelManager(object):
         try:
             # By passing the specified VAE to the conversion function, the autoencoder
             # will be built into the model rather than tacked on afterward via the config file
-            vae_model = self._load_vae(vae) if vae else None
+            vae_model=None
+            if vae:
+                vae_model=self._load_vae(vae)
+                vae_path=None
             convert_ckpt_to_diffusers(
                 ckpt_path,
                 diffusers_path,
                 extract_ema=True,
                 original_config_file=original_config_file,
                 vae=vae_model,
-                vae_path=str(vae_path) if vae_path else None,
+                vae_path=vae_path,
                 scan_needed=scan_needed,
             )
             print(
@@ -1060,36 +952,6 @@ class ModelManager(object):
 
         return search_folder, found_models
 
-    def _choose_diffusers_vae(
-        self, model_name: str, vae: str = None
-    ) -> Union[dict, str]:
-        # In the event that the original entry is using a custom ckpt VAE, we try to
-        # map that VAE onto a diffuser VAE using a hard-coded dictionary.
-        # I would prefer to do this differently: We load the ckpt model into memory, swap the
-        # VAE in memory, and then pass that to convert_ckpt_to_diffusers() so that the swapped
-        # VAE is built into the model. However, when I tried this I got obscure key errors.
-        if vae:
-            return vae
-        if model_name in self.config and (
-            vae_ckpt_path := self.model_info(model_name).get("vae", None)
-        ):
-            vae_basename = Path(vae_ckpt_path).stem
-            diffusers_vae = None
-            if diffusers_vae := VAE_TO_REPO_ID.get(vae_basename, None):
-                print(
-                    f">> {vae_basename} VAE corresponds to known {diffusers_vae} diffusers version"
-                )
-                vae = {"repo_id": diffusers_vae}
-            else:
-                print(
-                    f'** Custom VAE "{vae_basename}" found, but corresponding diffusers model unknown'
-                )
-                print(
-                    '** Using "stabilityai/sd-vae-ft-mse"; If this isn\'t right, please edit the model config'
-                )
-                vae = {"repo_id": "stabilityai/sd-vae-ft-mse"}
-        return vae
-
     def _make_cache_room(self) -> None:
         num_loaded_models = len(self.models)
         if num_loaded_models >= self.max_loaded_models:
@@ -1351,6 +1213,22 @@ class ModelManager(object):
             f.write(hash)
         return hash
 
+    @classmethod
+    def _scan_for_matching_file(
+            self,model_path: Path,
+            suffixes: List[str]=['.vae.pt','.vae.ckpt','.vae.safetensors']
+    )->Path:
+        """
+        Find a file with same basename as the indicated model, but with one
+        of the suffixes passed.
+        """
+        # look for a custom vae
+        vae_path = None
+        for suffix in suffixes:
+            if model_path.with_suffix(suffix).exists():
+                vae_path = model_path.with_suffix(suffix)
+        return vae_path
+                               
     def _load_vae(self, vae_config) -> AutoencoderKL:
         vae_args = {}
         try:
@@ -1362,7 +1240,7 @@ class ModelManager(object):
         using_fp16 = self.precision == "float16"
 
         vae_args.update(
-            cache_dir=global_cache_dir("hug"),
+            cache_dir=global_cache_dir("hub"),
             local_files_only=not Globals.internet_available,
         )
 

ldm/invoke/readline.py

@@ -11,9 +11,11 @@ seeds:
 import os
 import re
 import atexit
+from typing import List
 from ldm.invoke.args import Args
 from ldm.invoke.concepts_lib import HuggingFaceConceptsLibrary
 from ldm.invoke.globals import Globals
+from ldm.modules.lora_manager import LoraManager
 
 # ---------------readline utilities---------------------
 try:
@@ -136,6 +138,9 @@ class Completer(object):
             elif re.search('<[\w-]*$',buffer):
                 self.matches= self._concept_completions(text,state)
 
+            elif re.search('withLora\(?[a-zA-Z0-9._-]*$',buffer):
+                self.matches= self._lora_completions(text,state)
+
             # looking for a model
             elif re.match('^'+'|'.join(MODEL_COMMANDS),buffer):
                 self.matches= self._model_completions(text, state)
@@ -298,6 +303,15 @@ class Completer(object):
         matches.sort()
         return matches
 
+    def _lora_completions(self, text, state)->List[str]:
+        loras: dict = LoraManager.list_loras()
+        lora_names = [f'withLora({x},1)' for x in loras.keys()]
+        matches = list()
+        for lora in lora_names:
+            if lora.startswith(text):
+                matches.append(lora)
+        return sorted(matches)
+
     def _model_completions(self, text, state, ckpt_only=False):
         m = re.search('(!switch\s+)(\w*)',text)
         if m:

ldm/models/diffusion/cross_attention_control.py

@@ -288,16 +288,7 @@ class InvokeAICrossAttentionMixin:
         return self.einsum_op_tensor_mem(q, k, v, 32)
 
 
-
-def restore_default_cross_attention(model, is_running_diffusers: bool, restore_attention_processor: Optional[AttnProcessor]=None):
-    if is_running_diffusers:
-        unet = model
-        unet.set_attn_processor(restore_attention_processor or CrossAttnProcessor())
-    else:
-        remove_attention_function(model)
-
-
-def override_cross_attention(model, context: Context, is_running_diffusers = False):
+def setup_cross_attention_control_attention_processors(unet: UNet2DConditionModel, context: Context):
     """
     Inject attention parameters and functions into the passed in model to enable cross attention editing.
 
@@ -323,24 +314,15 @@ def override_cross_attention(model, context: Context, is_running_diffusers = Fal
 
     context.cross_attention_mask = mask.to(device)
     context.cross_attention_index_map = indices.to(device)
-    if is_running_diffusers:
-        unet = model
-        old_attn_processors = unet.attn_processors
-        if torch.backends.mps.is_available():
-            # see note in StableDiffusionGeneratorPipeline.__init__ about borked slicing on MPS
-            unet.set_attn_processor(SwapCrossAttnProcessor())
-        else:
-            # try to re-use an existing slice size
-            default_slice_size = 4
-            slice_size = next((p.slice_size for p in old_attn_processors.values() if type(p) is SlicedAttnProcessor), default_slice_size)
-            unet.set_attn_processor(SlicedSwapCrossAttnProcesser(slice_size=slice_size))
-        return old_attn_processors
+    old_attn_processors = unet.attn_processors
+    if torch.backends.mps.is_available():
+        # see note in StableDiffusionGeneratorPipeline.__init__ about borked slicing on MPS
+        unet.set_attn_processor(SwapCrossAttnProcessor())
     else:
-        context.register_cross_attention_modules(model)
-        inject_attention_function(model, context)
-        return None
-
-
+        # try to re-use an existing slice size
+        default_slice_size = 4
+        slice_size = next((p.slice_size for p in old_attn_processors.values() if type(p) is SlicedAttnProcessor), default_slice_size)
+        unet.set_attn_processor(SlicedSwapCrossAttnProcesser(slice_size=slice_size))
 
 
 def get_cross_attention_modules(model, which: CrossAttentionType) -> list[tuple[str, InvokeAICrossAttentionMixin]]:

ldm/models/diffusion/ddim.py

@@ -12,17 +12,6 @@ class DDIMSampler(Sampler):
         self.invokeai_diffuser = InvokeAIDiffuserComponent(self.model,
                                                            model_forward_callback = lambda x, sigma, cond: self.model.apply_model(x, sigma, cond))
 
-    def prepare_to_sample(self, t_enc, **kwargs):
-        super().prepare_to_sample(t_enc, **kwargs)
-
-        extra_conditioning_info = kwargs.get('extra_conditioning_info', None)
-        all_timesteps_count = kwargs.get('all_timesteps_count', t_enc)
-
-        if extra_conditioning_info is not None and extra_conditioning_info.wants_cross_attention_control:
-            self.invokeai_diffuser.override_cross_attention(extra_conditioning_info, step_count = all_timesteps_count)
-        else:
-            self.invokeai_diffuser.restore_default_cross_attention()
-
 
     # This is the central routine
     @torch.no_grad()

ldm/models/diffusion/ddpm.py

@@ -19,7 +19,7 @@ from functools import partial
 from tqdm import tqdm
 from torchvision.utils import make_grid
 from pytorch_lightning.utilities.distributed import rank_zero_only
-from omegaconf import ListConfig
+from omegaconf import ListConfig, OmegaConf
 import urllib
 
 from ldm.modules.textual_inversion_manager import TextualInversionManager
@@ -609,6 +609,7 @@ class DDPM(pl.LightningModule):
         opt = torch.optim.AdamW(params, lr=lr)
         return opt
 
+    
 
 class LatentDiffusion(DDPM):
     """main class"""
@@ -617,7 +618,7 @@ class LatentDiffusion(DDPM):
         self,
         first_stage_config,
         cond_stage_config,
-        personalization_config,
+        personalization_config=None,
         num_timesteps_cond=None,
         cond_stage_key='image',
         cond_stage_trainable=False,
@@ -675,7 +676,8 @@ class LatentDiffusion(DDPM):
         self.model.train = disabled_train
         for param in self.model.parameters():
             param.requires_grad = False
-
+            
+        personalization_config = personalization_config or self._fallback_personalization_config()
         self.embedding_manager = self.instantiate_embedding_manager(
             personalization_config, self.cond_stage_model
         )
@@ -2150,6 +2152,25 @@ class LatentDiffusion(DDPM):
 
                 self.emb_ckpt_counter += 500
 
+    @classmethod
+    def _fallback_personalization_config(self)->dict:
+        """
+        This protects us against custom legacy config files that
+        don't contain the personalization_config section.
+        """
+        return OmegaConf.create(
+            dict(
+                target='ldm.modules.embedding_manager.EmbeddingManager',
+                params=dict(
+                    placeholder_strings=list('*'),
+                    initializer_words=list('sculpture'),
+                    per_image_tokens=False,
+                    num_vectors_per_token=1,
+                    progressive_words=False,
+                )
+            )
+        )
+
 
 class DiffusionWrapper(pl.LightningModule):
     def __init__(self, diff_model_config, conditioning_key):

ldm/models/diffusion/ksampler.py

@@ -38,15 +38,6 @@ class CFGDenoiser(nn.Module):
                                                            model_forward_callback=lambda x, sigma, cond: self.inner_model(x, sigma, cond=cond))
 
 
-    def prepare_to_sample(self, t_enc, **kwargs):
-
-        extra_conditioning_info = kwargs.get('extra_conditioning_info', None)
-
-        if extra_conditioning_info is not None and extra_conditioning_info.wants_cross_attention_control:
-            self.invokeai_diffuser.override_cross_attention(extra_conditioning_info, step_count = t_enc)
-        else:
-            self.invokeai_diffuser.restore_default_cross_attention()
-
 
     def forward(self, x, sigma, uncond, cond, cond_scale):
         next_x = self.invokeai_diffuser.do_diffusion_step(x, sigma, uncond, cond, cond_scale)

ldm/models/diffusion/plms.py

@@ -14,17 +14,6 @@ class PLMSSampler(Sampler):
     def __init__(self, model, schedule='linear', device=None, **kwargs):
         super().__init__(model,schedule,model.num_timesteps, device)
 
-    def prepare_to_sample(self, t_enc, **kwargs):
-        super().prepare_to_sample(t_enc, **kwargs)
-
-        extra_conditioning_info = kwargs.get('extra_conditioning_info', None)
-        all_timesteps_count = kwargs.get('all_timesteps_count', t_enc)
-
-        if extra_conditioning_info is not None and extra_conditioning_info.wants_cross_attention_control:
-            self.invokeai_diffuser.override_cross_attention(extra_conditioning_info, step_count = all_timesteps_count)
-        else:
-            self.invokeai_diffuser.restore_default_cross_attention()
-
 
     # this is the essential routine
     @torch.no_grad()

ldm/models/diffusion/shared_invokeai_diffusion.py

@@ -1,25 +1,36 @@
 from contextlib import contextmanager
 from dataclasses import dataclass
 from math import ceil
-from typing import Callable, Optional, Union, Any, Dict
+from typing import Callable, Optional, Union, Any
 
 import numpy as np
 import torch
-from diffusers.models.cross_attention import AttnProcessor
+
+from diffusers import UNet2DConditionModel
 from typing_extensions import TypeAlias
 
 from ldm.invoke.globals import Globals
-from ldm.models.diffusion.cross_attention_control import Arguments, \
-    restore_default_cross_attention, override_cross_attention, Context, get_cross_attention_modules, \
-    CrossAttentionType, SwapCrossAttnContext
+from ldm.models.diffusion.cross_attention_control import (
+    Arguments,
+    setup_cross_attention_control_attention_processors,
+    Context,
+    get_cross_attention_modules,
+    CrossAttentionType,
+    SwapCrossAttnContext,
+)
 from ldm.models.diffusion.cross_attention_map_saving import AttentionMapSaver
+from ldm.modules.lora_manager import LoraCondition
 
 ModelForwardCallback: TypeAlias = Union[
     # x, t, conditioning, Optional[cross-attention kwargs]
-    Callable[[torch.Tensor, torch.Tensor, torch.Tensor, Optional[dict[str, Any]]], torch.Tensor],
-    Callable[[torch.Tensor, torch.Tensor, torch.Tensor], torch.Tensor]
+    Callable[
+        [torch.Tensor, torch.Tensor, torch.Tensor, Optional[dict[str, Any]]],
+        torch.Tensor,
+    ],
+    Callable[[torch.Tensor, torch.Tensor, torch.Tensor], torch.Tensor],
 ]
 
+
 @dataclass(frozen=True)
 class PostprocessingSettings:
     threshold: float
@@ -29,31 +40,39 @@ class PostprocessingSettings:
 
 
 class InvokeAIDiffuserComponent:
-    '''
+    """
     The aim of this component is to provide a single place for code that can be applied identically to
     all InvokeAI diffusion procedures.
 
     At the moment it includes the following features:
     * Cross attention control ("prompt2prompt")
     * Hybrid conditioning (used for inpainting)
-    '''
+    * "LoRA" and "PEFT" augmentions to the unet's attention weights
+    """
+
     debug_thresholding = False
     sequential_guidance = False
 
     @dataclass
     class ExtraConditioningInfo:
-
         tokens_count_including_eos_bos: int
         cross_attention_control_args: Optional[Arguments] = None
+        lora_conditions: Optional[list[LoraCondition]] = None
 
         @property
         def wants_cross_attention_control(self):
             return self.cross_attention_control_args is not None
 
+        @property
+        def has_lora_conditions(self):
+            return self.lora_conditions is not None
 
-    def __init__(self, model, model_forward_callback: ModelForwardCallback,
-                 is_running_diffusers: bool=False,
-                 ):
+    def __init__(
+        self,
+        model,
+        model_forward_callback: ModelForwardCallback,
+        is_running_diffusers: bool = False,
+    ):
         """
         :param model: the unet model to pass through to cross attention control
         :param model_forward_callback: a lambda with arguments (x, sigma, conditioning_to_apply). will be called repeatedly. most likely, this should simply call model.forward(x, sigma, conditioning)
@@ -65,43 +84,44 @@ class InvokeAIDiffuserComponent:
         self.cross_attention_control_context = None
         self.sequential_guidance = Globals.sequential_guidance
 
+    @classmethod
     @contextmanager
-    def custom_attention_context(self,
-                                 extra_conditioning_info: Optional[ExtraConditioningInfo],
-                                 step_count: int):
-        do_swap = extra_conditioning_info is not None and extra_conditioning_info.wants_cross_attention_control
-        old_attn_processor = None
-        if do_swap:
-            old_attn_processor = self.override_cross_attention(extra_conditioning_info,
-                                                               step_count=step_count)
+    def custom_attention_context(
+        clss,
+        unet: UNet2DConditionModel, # note: also may futz with the text encoder depending on requested LoRAs
+        extra_conditioning_info: Optional[ExtraConditioningInfo],
+        step_count: int
+    ):
+        old_attn_processors = None
+        if extra_conditioning_info and (
+            extra_conditioning_info.wants_cross_attention_control
+            | extra_conditioning_info.has_lora_conditions
+        ):
+            old_attn_processors = unet.attn_processors
+            # Load lora conditions into the model
+            if extra_conditioning_info.has_lora_conditions:
+                for condition in extra_conditioning_info.lora_conditions:
+                    condition() # target model is stored in condition state for some reason
+            if extra_conditioning_info.wants_cross_attention_control:
+                cross_attention_control_context = Context(
+                    arguments=extra_conditioning_info.cross_attention_control_args,
+                    step_count=step_count,
+                )
+                setup_cross_attention_control_attention_processors(
+                    unet,
+                    cross_attention_control_context,
+                )
+
         try:
             yield None
         finally:
-            if old_attn_processor is not None:
-                self.restore_default_cross_attention(old_attn_processor)
+            if old_attn_processors is not None:
+                unet.set_attn_processor(old_attn_processors)
+            if extra_conditioning_info and extra_conditioning_info.has_lora_conditions:
+                for lora_condition in extra_conditioning_info.lora_conditions:
+                    lora_condition.unload()
             # TODO resuscitate attention map saving
-            #self.remove_attention_map_saving()
-
-    def override_cross_attention(self, conditioning: ExtraConditioningInfo, step_count: int) -> Dict[str, AttnProcessor]:
-        """
-        setup cross attention .swap control. for diffusers this replaces the attention processor, so
-        the previous attention processor is returned so that the caller can restore it later.
-        """
-        self.conditioning = conditioning
-        self.cross_attention_control_context = Context(
-            arguments=self.conditioning.cross_attention_control_args,
-            step_count=step_count
-        )
-        return override_cross_attention(self.model,
-                                        self.cross_attention_control_context,
-                                        is_running_diffusers=self.is_running_diffusers)
-
-    def restore_default_cross_attention(self, restore_attention_processor: Optional['AttnProcessor']=None):
-        self.conditioning = None
-        self.cross_attention_control_context = None
-        restore_default_cross_attention(self.model,
-                                        is_running_diffusers=self.is_running_diffusers,
-                                        restore_attention_processor=restore_attention_processor)
+            # self.remove_attention_map_saving()
 
     def setup_attention_map_saving(self, saver: AttentionMapSaver):
         def callback(slice, dim, offset, slice_size, key):
@@ -110,26 +130,40 @@ class InvokeAIDiffuserComponent:
                 return
             saver.add_attention_maps(slice, key)
 
-        tokens_cross_attention_modules = get_cross_attention_modules(self.model, CrossAttentionType.TOKENS)
+        tokens_cross_attention_modules = get_cross_attention_modules(
+            self.model, CrossAttentionType.TOKENS
+        )
         for identifier, module in tokens_cross_attention_modules:
-            key = ('down' if identifier.startswith('down') else
-                   'up' if identifier.startswith('up') else
-                   'mid')
+            key = (
+                "down"
+                if identifier.startswith("down")
+                else "up"
+                if identifier.startswith("up")
+                else "mid"
+            )
             module.set_attention_slice_calculated_callback(
-                lambda slice, dim, offset, slice_size, key=key: callback(slice, dim, offset, slice_size, key))
+                lambda slice, dim, offset, slice_size, key=key: callback(
+                    slice, dim, offset, slice_size, key
+                )
+            )
 
     def remove_attention_map_saving(self):
-        tokens_cross_attention_modules = get_cross_attention_modules(self.model, CrossAttentionType.TOKENS)
+        tokens_cross_attention_modules = get_cross_attention_modules(
+            self.model, CrossAttentionType.TOKENS
+        )
         for _, module in tokens_cross_attention_modules:
             module.set_attention_slice_calculated_callback(None)
 
-    def do_diffusion_step(self, x: torch.Tensor, sigma: torch.Tensor,
-                                unconditioning: Union[torch.Tensor,dict],
-                                conditioning: Union[torch.Tensor,dict],
-                                unconditional_guidance_scale: float,
-                                step_index: Optional[int]=None,
-                                total_step_count: Optional[int]=None,
-                          ):
+    def do_diffusion_step(
+        self,
+        x: torch.Tensor,
+        sigma: torch.Tensor,
+        unconditioning: Union[torch.Tensor, dict],
+        conditioning: Union[torch.Tensor, dict],
+        unconditional_guidance_scale: float,
+        step_index: Optional[int] = None,
+        total_step_count: Optional[int] = None,
+    ):
         """
         :param x: current latents
         :param sigma: aka t, passed to the internal model to control how much denoising will occur
@@ -140,33 +174,55 @@ class InvokeAIDiffuserComponent:
         :return: the new latents after applying the model to x using unscaled unconditioning and CFG-scaled conditioning.
         """
 
-
         cross_attention_control_types_to_do = []
         context: Context = self.cross_attention_control_context
         if self.cross_attention_control_context is not None:
-            percent_through = self.calculate_percent_through(sigma, step_index, total_step_count)
-            cross_attention_control_types_to_do = context.get_active_cross_attention_control_types_for_step(percent_through)
+            percent_through = self.calculate_percent_through(
+                sigma, step_index, total_step_count
+            )
+            cross_attention_control_types_to_do = (
+                context.get_active_cross_attention_control_types_for_step(
+                    percent_through
+                )
+            )
 
-        wants_cross_attention_control = (len(cross_attention_control_types_to_do) > 0)
+        wants_cross_attention_control = len(cross_attention_control_types_to_do) > 0
         wants_hybrid_conditioning = isinstance(conditioning, dict)
 
         if wants_hybrid_conditioning:
-            unconditioned_next_x, conditioned_next_x = self._apply_hybrid_conditioning(x, sigma, unconditioning,
-                                                                                       conditioning)
+            unconditioned_next_x, conditioned_next_x = self._apply_hybrid_conditioning(
+                x, sigma, unconditioning, conditioning
+            )
         elif wants_cross_attention_control:
-            unconditioned_next_x, conditioned_next_x = self._apply_cross_attention_controlled_conditioning(x, sigma,
-                                                                                                           unconditioning,
-                                                                                                           conditioning,
-                                                                                                           cross_attention_control_types_to_do)
+            (
+                unconditioned_next_x,
+                conditioned_next_x,
+            ) = self._apply_cross_attention_controlled_conditioning(
+                x,
+                sigma,
+                unconditioning,
+                conditioning,
+                cross_attention_control_types_to_do,
+            )
         elif self.sequential_guidance:
-            unconditioned_next_x, conditioned_next_x = self._apply_standard_conditioning_sequentially(
-                x, sigma, unconditioning, conditioning)
+            (
+                unconditioned_next_x,
+                conditioned_next_x,
+            ) = self._apply_standard_conditioning_sequentially(
+                x, sigma, unconditioning, conditioning
+            )
 
         else:
-            unconditioned_next_x, conditioned_next_x = self._apply_standard_conditioning(
-                x, sigma, unconditioning, conditioning)
+            (
+                unconditioned_next_x,
+                conditioned_next_x,
+            ) = self._apply_standard_conditioning(
+                x, sigma, unconditioning, conditioning
+            )
 
-        combined_next_x = self._combine(unconditioned_next_x, conditioned_next_x, unconditional_guidance_scale)
+        combined_next_x = self._combine(
+            unconditioned_next_x, conditioned_next_x, unconditional_guidance_scale
+        )
 
         return combined_next_x
 
@@ -176,24 +232,33 @@ class InvokeAIDiffuserComponent:
         latents: torch.Tensor,
         sigma,
         step_index,
-        total_step_count
+        total_step_count,
     ) -> torch.Tensor:
         if postprocessing_settings is not None:
-            percent_through = self.calculate_percent_through(sigma, step_index, total_step_count)
-            latents = self.apply_threshold(postprocessing_settings, latents, percent_through)
-            latents = self.apply_symmetry(postprocessing_settings, latents, percent_through)
+            percent_through = self.calculate_percent_through(
+                sigma, step_index, total_step_count
+            )
+            latents = self.apply_threshold(
+                postprocessing_settings, latents, percent_through
+            )
+            latents = self.apply_symmetry(
+                postprocessing_settings, latents, percent_through
+            )
         return latents
 
     def calculate_percent_through(self, sigma, step_index, total_step_count):
         if step_index is not None and total_step_count is not None:
             # 🧨diffusers codepath
-            percent_through = step_index / total_step_count  # will never reach 1.0 - this is deliberate
+            percent_through = (
+                step_index / total_step_count
+            )  # will never reach 1.0 - this is deliberate
         else:
             # legacy compvis codepath
             # TODO remove when compvis codepath support is dropped
             if step_index is None and sigma is None:
                 raise ValueError(
-                    f"Either step_index or sigma is required when doing cross attention control, but both are None.")
+                    f"Either step_index or sigma is required when doing cross attention control, but both are None."
+                )
             percent_through = self.estimate_percent_through(step_index, sigma)
         return percent_through
 
@@ -204,24 +269,30 @@ class InvokeAIDiffuserComponent:
         x_twice = torch.cat([x] * 2)
         sigma_twice = torch.cat([sigma] * 2)
         both_conditionings = torch.cat([unconditioning, conditioning])
-        both_results = self.model_forward_callback(x_twice, sigma_twice, both_conditionings)
+        both_results = self.model_forward_callback(
+            x_twice, sigma_twice, both_conditionings
+        )
         unconditioned_next_x, conditioned_next_x = both_results.chunk(2)
-        if conditioned_next_x.device.type == 'mps':
+        if conditioned_next_x.device.type == "mps":
             # prevent a result filled with zeros. seems to be a torch bug.
             conditioned_next_x = conditioned_next_x.clone()
         return unconditioned_next_x, conditioned_next_x
 
-
-    def _apply_standard_conditioning_sequentially(self, x: torch.Tensor, sigma, unconditioning: torch.Tensor, conditioning: torch.Tensor):
+    def _apply_standard_conditioning_sequentially(
+        self,
+        x: torch.Tensor,
+        sigma,
+        unconditioning: torch.Tensor,
+        conditioning: torch.Tensor,
+    ):
         # low-memory sequential path
         unconditioned_next_x = self.model_forward_callback(x, sigma, unconditioning)
         conditioned_next_x = self.model_forward_callback(x, sigma, conditioning)
-        if conditioned_next_x.device.type == 'mps':
+        if conditioned_next_x.device.type == "mps":
             # prevent a result filled with zeros. seems to be a torch bug.
             conditioned_next_x = conditioned_next_x.clone()
         return unconditioned_next_x, conditioned_next_x
 
-
     def _apply_hybrid_conditioning(self, x, sigma, unconditioning, conditioning):
         assert isinstance(conditioning, dict)
         assert isinstance(unconditioning, dict)
@@ -236,48 +307,80 @@ class InvokeAIDiffuserComponent:
                 ]
             else:
                 both_conditionings[k] = torch.cat([unconditioning[k], conditioning[k]])
-        unconditioned_next_x, conditioned_next_x = self.model_forward_callback(x_twice, sigma_twice, both_conditionings).chunk(2)
+        unconditioned_next_x, conditioned_next_x = self.model_forward_callback(
+            x_twice, sigma_twice, both_conditionings
+        ).chunk(2)
         return unconditioned_next_x, conditioned_next_x
 
-
-    def _apply_cross_attention_controlled_conditioning(self,
-                                                     x: torch.Tensor,
-                                                     sigma,
-                                                     unconditioning,
-                                                     conditioning,
-                                                     cross_attention_control_types_to_do):
+    def _apply_cross_attention_controlled_conditioning(
+        self,
+        x: torch.Tensor,
+        sigma,
+        unconditioning,
+        conditioning,
+        cross_attention_control_types_to_do,
+    ):
         if self.is_running_diffusers:
-            return self._apply_cross_attention_controlled_conditioning__diffusers(x, sigma, unconditioning,
-                                                                                  conditioning,
-                                                                                  cross_attention_control_types_to_do)
+            return self._apply_cross_attention_controlled_conditioning__diffusers(
+                x,
+                sigma,
+                unconditioning,
+                conditioning,
+                cross_attention_control_types_to_do,
+            )
         else:
-            return self._apply_cross_attention_controlled_conditioning__compvis(x, sigma, unconditioning, conditioning,
-                                                                                cross_attention_control_types_to_do)
+            return self._apply_cross_attention_controlled_conditioning__compvis(
+                x,
+                sigma,
+                unconditioning,
+                conditioning,
+                cross_attention_control_types_to_do,
+            )
 
-    def _apply_cross_attention_controlled_conditioning__diffusers(self,
-                                                                 x: torch.Tensor,
-                                                                 sigma,
-                                                                 unconditioning,
-                                                                 conditioning,
-                                                                 cross_attention_control_types_to_do):
+    def _apply_cross_attention_controlled_conditioning__diffusers(
+        self,
+        x: torch.Tensor,
+        sigma,
+        unconditioning,
+        conditioning,
+        cross_attention_control_types_to_do,
+    ):
         context: Context = self.cross_attention_control_context
 
-        cross_attn_processor_context = SwapCrossAttnContext(modified_text_embeddings=context.arguments.edited_conditioning,
-                                                            index_map=context.cross_attention_index_map,
-                                                            mask=context.cross_attention_mask,
-                                                            cross_attention_types_to_do=[])
+        cross_attn_processor_context = SwapCrossAttnContext(
+            modified_text_embeddings=context.arguments.edited_conditioning,
+            index_map=context.cross_attention_index_map,
+            mask=context.cross_attention_mask,
+            cross_attention_types_to_do=[],
+        )
         # no cross attention for unconditioning (negative prompt)
-        unconditioned_next_x = self.model_forward_callback(x, sigma, unconditioning,
-                                                           {"swap_cross_attn_context": cross_attn_processor_context})
+        unconditioned_next_x = self.model_forward_callback(
+            x,
+            sigma,
+            unconditioning,
+            {"swap_cross_attn_context": cross_attn_processor_context},
+        )
 
         # do requested cross attention types for conditioning (positive prompt)
-        cross_attn_processor_context.cross_attention_types_to_do = cross_attention_control_types_to_do
-        conditioned_next_x = self.model_forward_callback(x, sigma, conditioning,
-                                                         {"swap_cross_attn_context": cross_attn_processor_context})
+        cross_attn_processor_context.cross_attention_types_to_do = (
+            cross_attention_control_types_to_do
+        )
+        conditioned_next_x = self.model_forward_callback(
+            x,
+            sigma,
+            conditioning,
+            {"swap_cross_attn_context": cross_attn_processor_context},
+        )
         return unconditioned_next_x, conditioned_next_x
 
-
-    def _apply_cross_attention_controlled_conditioning__compvis(self, x:torch.Tensor, sigma, unconditioning, conditioning, cross_attention_control_types_to_do):
+    def _apply_cross_attention_controlled_conditioning__compvis(
+        self,
+        x: torch.Tensor,
+        sigma,
+        unconditioning,
+        conditioning,
+        cross_attention_control_types_to_do,
+    ):
         # print('pct', percent_through, ': doing cross attention control on', cross_attention_control_types_to_do)
         # slower non-batched path (20% slower on mac MPS)
         # We are only interested in using attention maps for conditioned_next_x, but batching them with generation of
@@ -287,24 +390,26 @@ class InvokeAIDiffuserComponent:
         # representing batched uncond + cond, but then when it comes to applying the saved attention, the
         # wrangler gets an attention tensor which only has shape[0]=8, representing just self.edited_conditionings.)
         # todo: give CrossAttentionControl's `wrangler` function more info so it can work with a batched call as well.
-        context:Context = self.cross_attention_control_context
+        context: Context = self.cross_attention_control_context
 
         try:
             unconditioned_next_x = self.model_forward_callback(x, sigma, unconditioning)
 
             # process x using the original prompt, saving the attention maps
-            #print("saving attention maps for", cross_attention_control_types_to_do)
+            # print("saving attention maps for", cross_attention_control_types_to_do)
             for ca_type in cross_attention_control_types_to_do:
                 context.request_save_attention_maps(ca_type)
             _ = self.model_forward_callback(x, sigma, conditioning)
             context.clear_requests(cleanup=False)
 
             # process x again, using the saved attention maps to control where self.edited_conditioning will be applied
-            #print("applying saved attention maps for", cross_attention_control_types_to_do)
+            # print("applying saved attention maps for", cross_attention_control_types_to_do)
             for ca_type in cross_attention_control_types_to_do:
                 context.request_apply_saved_attention_maps(ca_type)
-            edited_conditioning = self.conditioning.cross_attention_control_args.edited_conditioning
-            conditioned_next_x = self.model_forward_callback(x, sigma, edited_conditioning)
+            edited_conditioning = context.arguments.edited_conditioning
+            conditioned_next_x = self.model_forward_callback(
+                x, sigma, edited_conditioning
+            )
             context.clear_requests(cleanup=True)
 
         except:
@@ -323,17 +428,21 @@ class InvokeAIDiffuserComponent:
         self,
         postprocessing_settings: PostprocessingSettings,
         latents: torch.Tensor,
-        percent_through: float
+        percent_through: float,
     ) -> torch.Tensor:
-
-        if postprocessing_settings.threshold is None or postprocessing_settings.threshold == 0.0:
+        if (
+            postprocessing_settings.threshold is None
+            or postprocessing_settings.threshold == 0.0
+        ):
             return latents
 
         threshold = postprocessing_settings.threshold
         warmup = postprocessing_settings.warmup
 
         if percent_through < warmup:
-            current_threshold = threshold + threshold * 5 * (1 - (percent_through / warmup))
+            current_threshold = threshold + threshold * 5 * (
+                1 - (percent_through / warmup)
+            )
         else:
             current_threshold = threshold
 
@@ -347,10 +456,14 @@ class InvokeAIDiffuserComponent:
 
         if self.debug_thresholding:
             std, mean = [i.item() for i in torch.std_mean(latents)]
-            outside = torch.count_nonzero((latents < -current_threshold) | (latents > current_threshold))
-            print(f"\nThreshold: %={percent_through} threshold={current_threshold:.3f} (of {threshold:.3f})\n"
-                  f"  | min, mean, max = {minval:.3f}, {mean:.3f}, {maxval:.3f}\tstd={std}\n"
-                  f"  | {outside / latents.numel() * 100:.2f}% values outside threshold")
+            outside = torch.count_nonzero(
+                (latents < -current_threshold) | (latents > current_threshold)
+            )
+            print(
+                f"\nThreshold: %={percent_through} threshold={current_threshold:.3f} (of {threshold:.3f})\n"
+                f"  | min, mean, max = {minval:.3f}, {mean:.3f}, {maxval:.3f}\tstd={std}\n"
+                f"  | {outside / latents.numel() * 100:.2f}% values outside threshold"
+            )
 
         if maxval < current_threshold and minval > -current_threshold:
             return latents
@@ -363,17 +476,23 @@ class InvokeAIDiffuserComponent:
             latents = torch.clone(latents)
             maxval = np.clip(maxval * scale, 1, current_threshold)
             num_altered += torch.count_nonzero(latents > maxval)
-            latents[latents > maxval] = torch.rand_like(latents[latents > maxval]) * maxval
+            latents[latents > maxval] = (
+                torch.rand_like(latents[latents > maxval]) * maxval
+            )
 
         if minval < -current_threshold:
             latents = torch.clone(latents)
             minval = np.clip(minval * scale, -current_threshold, -1)
             num_altered += torch.count_nonzero(latents < minval)
-            latents[latents < minval] = torch.rand_like(latents[latents < minval]) * minval
+            latents[latents < minval] = (
+                torch.rand_like(latents[latents < minval]) * minval
+            )
 
         if self.debug_thresholding:
-            print(f"  | min,     , max = {minval:.3f},        , {maxval:.3f}\t(scaled by {scale})\n"
-                  f"  | {num_altered / latents.numel() * 100:.2f}% values altered")
+            print(
+                f"  | min,     , max = {minval:.3f},        , {maxval:.3f}\t(scaled by {scale})\n"
+                f"  | {num_altered / latents.numel() * 100:.2f}% values altered"
+            )
 
         return latents
 
@@ -381,9 +500,8 @@ class InvokeAIDiffuserComponent:
         self,
         postprocessing_settings: PostprocessingSettings,
         latents: torch.Tensor,
-        percent_through: float
+        percent_through: float,
     ) -> torch.Tensor:
-
         # Reset our last percent through if this is our first step.
         if percent_through == 0.0:
             self.last_percent_through = 0.0
@@ -393,36 +511,52 @@ class InvokeAIDiffuserComponent:
 
         # Check for out of bounds
         h_symmetry_time_pct = postprocessing_settings.h_symmetry_time_pct
-        if (h_symmetry_time_pct is not None and (h_symmetry_time_pct <= 0.0 or h_symmetry_time_pct > 1.0)):
+        if h_symmetry_time_pct is not None and (
+            h_symmetry_time_pct <= 0.0 or h_symmetry_time_pct > 1.0
+        ):
             h_symmetry_time_pct = None
 
         v_symmetry_time_pct = postprocessing_settings.v_symmetry_time_pct
-        if (v_symmetry_time_pct is not None and (v_symmetry_time_pct <= 0.0 or v_symmetry_time_pct > 1.0)):
+        if v_symmetry_time_pct is not None and (
+            v_symmetry_time_pct <= 0.0 or v_symmetry_time_pct > 1.0
+        ):
             v_symmetry_time_pct = None
 
         dev = latents.device.type
 
-        latents.to(device='cpu')
+        latents.to(device="cpu")
 
         if (
-            h_symmetry_time_pct != None and
-            self.last_percent_through < h_symmetry_time_pct and
-            percent_through >= h_symmetry_time_pct
+            h_symmetry_time_pct != None
+            and self.last_percent_through < h_symmetry_time_pct
+            and percent_through >= h_symmetry_time_pct
         ):
             # Horizontal symmetry occurs on the 3rd dimension of the latent
             width = latents.shape[3]
             x_flipped = torch.flip(latents, dims=[3])
-            latents = torch.cat([latents[:, :, :, 0:int(width/2)], x_flipped[:, :, :, int(width/2):int(width)]], dim=3)
+            latents = torch.cat(
+                [
+                    latents[:, :, :, 0 : int(width / 2)],
+                    x_flipped[:, :, :, int(width / 2) : int(width)],
+                ],
+                dim=3,
+            )
 
         if (
-            v_symmetry_time_pct != None and
-            self.last_percent_through < v_symmetry_time_pct and
-            percent_through >= v_symmetry_time_pct
+            v_symmetry_time_pct != None
+            and self.last_percent_through < v_symmetry_time_pct
+            and percent_through >= v_symmetry_time_pct
         ):
             # Vertical symmetry occurs on the 2nd dimension of the latent
             height = latents.shape[2]
             y_flipped = torch.flip(latents, dims=[2])
-            latents = torch.cat([latents[:, :, 0:int(height / 2)], y_flipped[:, :, int(height / 2):int(height)]], dim=2)
+            latents = torch.cat(
+                [
+                    latents[:, :, 0 : int(height / 2)],
+                    y_flipped[:, :, int(height / 2) : int(height)],
+                ],
+                dim=2,
+            )
 
         self.last_percent_through = percent_through
         return latents.to(device=dev)
@@ -430,7 +564,9 @@ class InvokeAIDiffuserComponent:
     def estimate_percent_through(self, step_index, sigma):
         if step_index is not None and self.cross_attention_control_context is not None:
             # percent_through will never reach 1.0 (but this is intended)
-            return float(step_index) / float(self.cross_attention_control_context.step_count)
+            return float(step_index) / float(
+                self.cross_attention_control_context.step_count
+            )
         # find the best possible index of the current sigma in the sigma sequence
         smaller_sigmas = torch.nonzero(self.model.sigmas <= sigma)
         sigma_index = smaller_sigmas[-1].item() if smaller_sigmas.shape[0] > 0 else 0
@@ -439,33 +575,38 @@ class InvokeAIDiffuserComponent:
         return 1.0 - float(sigma_index + 1) / float(self.model.sigmas.shape[0])
         # print('estimated percent_through', percent_through, 'from sigma', sigma.item())
 
-
     # todo: make this work
     @classmethod
-    def apply_conjunction(cls, x, t, forward_func, uc, c_or_weighted_c_list, global_guidance_scale):
+    def apply_conjunction(
+        cls, x, t, forward_func, uc, c_or_weighted_c_list, global_guidance_scale
+    ):
         x_in = torch.cat([x] * 2)
-        t_in = torch.cat([t] * 2) # aka sigmas
+        t_in = torch.cat([t] * 2)  # aka sigmas
 
         deltas = None
         uncond_latents = None
-        weighted_cond_list = c_or_weighted_c_list if type(c_or_weighted_c_list) is list else [(c_or_weighted_c_list, 1)]
+        weighted_cond_list = (
+            c_or_weighted_c_list
+            if type(c_or_weighted_c_list) is list
+            else [(c_or_weighted_c_list, 1)]
+        )
 
         # below is fugly omg
         num_actual_conditionings = len(c_or_weighted_c_list)
-        conditionings = [uc] + [c for c,weight in weighted_cond_list]
-        weights = [1] + [weight for c,weight in weighted_cond_list]
-        chunk_count = ceil(len(conditionings)/2)
+        conditionings = [uc] + [c for c, weight in weighted_cond_list]
+        weights = [1] + [weight for c, weight in weighted_cond_list]
+        chunk_count = ceil(len(conditionings) / 2)
         deltas = None
         for chunk_index in range(chunk_count):
-            offset = chunk_index*2
-            chunk_size = min(2, len(conditionings)-offset)
+            offset = chunk_index * 2
+            chunk_size = min(2, len(conditionings) - offset)
 
             if chunk_size == 1:
                 c_in = conditionings[offset]
                 latents_a = forward_func(x_in[:-1], t_in[:-1], c_in)
                 latents_b = None
             else:
-                c_in = torch.cat(conditionings[offset:offset+2])
+                c_in = torch.cat(conditionings[offset : offset + 2])
                 latents_a, latents_b = forward_func(x_in, t_in, c_in).chunk(2)
 
             # first chunk is guaranteed to be 2 entries: uncond_latents + first conditioining
@@ -478,11 +619,15 @@ class InvokeAIDiffuserComponent:
                     deltas = torch.cat((deltas, latents_b - uncond_latents))
 
         # merge the weighted deltas together into a single merged delta
-        per_delta_weights = torch.tensor(weights[1:], dtype=deltas.dtype, device=deltas.device)
+        per_delta_weights = torch.tensor(
+            weights[1:], dtype=deltas.dtype, device=deltas.device
+        )
         normalize = False
         if normalize:
             per_delta_weights /= torch.sum(per_delta_weights)
-        reshaped_weights = per_delta_weights.reshape(per_delta_weights.shape + (1, 1, 1))
+        reshaped_weights = per_delta_weights.reshape(
+            per_delta_weights.shape + (1, 1, 1)
+        )
         deltas_merged = torch.sum(deltas * reshaped_weights, dim=0, keepdim=True)
 
         # old_return_value = super().forward(x, sigma, uncond, cond, cond_scale)

ldm/modules/encoders/modules.py

@@ -463,6 +463,9 @@ class FrozenCLIPEmbedder(AbstractEncoder):
     def encode(self, text, **kwargs):
         return self(text, **kwargs)
 
+    def set_textual_inversion_manager(self, manager): #TextualInversionManager):
+        self.textual_inversion_manager = manager
+
     @property
     def device(self):
         return self.transformer.device
@@ -476,10 +479,6 @@ class WeightedFrozenCLIPEmbedder(FrozenCLIPEmbedder):
     fragment_weights_key = "fragment_weights"
     return_tokens_key = "return_tokens"
 
-    def set_textual_inversion_manager(self, manager): #TextualInversionManager):
-        # TODO all of the weighting and expanding stuff needs be moved out of this class
-        self.textual_inversion_manager = manager
-
     def forward(self, text: list, **kwargs):
         # TODO all of the weighting and expanding stuff needs be moved out of this class
         '''

ldm/modules/kohya_lora_manager.py

@@ -0,0 +1,589 @@
+import json
+from pathlib import Path
+from typing import Optional
+
+import torch
+from diffusers.models import UNet2DConditionModel
+from filelock import FileLock, Timeout
+from safetensors.torch import load_file
+from torch.utils.hooks import RemovableHandle
+from transformers import CLIPTextModel
+
+from ..invoke.globals import global_lora_models_dir
+from ..invoke.devices import choose_torch_device
+
+"""
+This module supports loading LoRA weights trained with https://github.com/kohya-ss/sd-scripts
+To be removed once support for diffusers LoRA weights is well supported
+"""
+
+
+class IncompatibleModelException(Exception):
+    "Raised when there is an attempt to load a LoRA into a model that is incompatible with it"
+    pass
+
+
+class LoRALayer:
+    lora_name: str
+    name: str
+    scale: float
+
+    up: torch.nn.Module
+    mid: Optional[torch.nn.Module] = None
+    down: torch.nn.Module
+
+    def __init__(self, lora_name: str, name: str, rank=4, alpha=1.0):
+        self.lora_name = lora_name
+        self.name = name
+        self.scale = alpha / rank if (alpha and rank) else 1.0
+
+    def forward(self, lora, input_h):
+        if self.mid is None:
+            weight = self.up(self.down(*input_h))
+        else:
+            weight = self.up(self.mid(self.down(*input_h)))
+
+        return weight * lora.multiplier * self.scale
+
+
+class LoHALayer:
+    lora_name: str
+    name: str
+    scale: float
+
+    w1_a: torch.Tensor
+    w1_b: torch.Tensor
+    w2_a: torch.Tensor
+    w2_b: torch.Tensor
+    t1: Optional[torch.Tensor] = None
+    t2: Optional[torch.Tensor] = None
+    bias: Optional[torch.Tensor] = None
+
+    org_module: torch.nn.Module
+
+    def __init__(self, lora_name: str, name: str, rank=4, alpha=1.0):
+        self.lora_name = lora_name
+        self.name = name
+        self.scale = alpha / rank if (alpha and rank) else 1.0
+
+    def forward(self, lora, input_h):
+        if type(self.org_module) == torch.nn.Conv2d:
+            op = torch.nn.functional.conv2d
+            extra_args = dict(
+                stride=self.org_module.stride,
+                padding=self.org_module.padding,
+                dilation=self.org_module.dilation,
+                groups=self.org_module.groups,
+            )
+
+        else:
+            op = torch.nn.functional.linear
+            extra_args = {}
+
+        if self.t1 is None:
+            weight = (self.w1_a @ self.w1_b) * (self.w2_a @ self.w2_b)
+
+        else:
+            rebuild1 = torch.einsum(
+                "i j k l, j r, i p -> p r k l", self.t1, self.w1_b, self.w1_a
+            )
+            rebuild2 = torch.einsum(
+                "i j k l, j r, i p -> p r k l", self.t2, self.w2_b, self.w2_a
+            )
+            weight = rebuild1 * rebuild2
+
+        bias = self.bias if self.bias is not None else 0
+        return op(
+            *input_h,
+            (weight + bias).view(self.org_module.weight.shape),
+            None,
+            **extra_args,
+        ) * lora.multiplier * self.scale
+
+class LoKRLayer:
+    lora_name: str
+    name: str
+    scale: float
+
+    w1: Optional[torch.Tensor] = None
+    w1_a: Optional[torch.Tensor] = None
+    w1_b: Optional[torch.Tensor] = None
+    w2: Optional[torch.Tensor] = None
+    w2_a: Optional[torch.Tensor] = None
+    w2_b: Optional[torch.Tensor] = None
+    t2: Optional[torch.Tensor] = None
+    bias: Optional[torch.Tensor] = None
+
+    org_module: torch.nn.Module
+
+    def __init__(self, lora_name: str, name: str, rank=4, alpha=1.0):
+        self.lora_name = lora_name
+        self.name = name
+        self.scale = alpha / rank if (alpha and rank) else 1.0
+
+    def forward(self, lora, input_h):
+
+        if type(self.org_module) == torch.nn.Conv2d:
+            op = torch.nn.functional.conv2d
+            extra_args = dict(
+                stride=self.org_module.stride,
+                padding=self.org_module.padding,
+                dilation=self.org_module.dilation,
+                groups=self.org_module.groups,
+            )
+
+        else:
+            op = torch.nn.functional.linear
+            extra_args = {}
+
+        w1 = self.w1
+        if w1 is None:
+            w1 = self.w1_a @ self.w1_b
+
+        w2 = self.w2
+        if w2 is None:
+            if self.t2 is None:
+                w2 = self.w2_a @ self.w2_b
+            else:
+                w2 = torch.einsum('i j k l, i p, j r -> p r k l', self.t2, self.w2_a, self.w2_b)
+
+
+        if len(w2.shape) == 4:
+            w1 = w1.unsqueeze(2).unsqueeze(2)
+        w2 = w2.contiguous()
+        weight = torch.kron(w1, w2).reshape(self.org_module.weight.shape)
+
+
+        bias = self.bias if self.bias is not None else 0
+        return op(
+            *input_h, 
+            (weight + bias).view(self.org_module.weight.shape),
+            None,
+            **extra_args
+        ) * lora.multiplier * self.scale
+
+
+class LoRAModuleWrapper:
+    unet: UNet2DConditionModel
+    text_encoder: CLIPTextModel
+    hooks: list[RemovableHandle]
+
+    def __init__(self, unet, text_encoder):
+        self.unet = unet
+        self.text_encoder = text_encoder
+        self.hooks = []
+        self.text_modules = None
+        self.unet_modules = None
+
+        self.applied_loras = {}
+        self.loaded_loras = {}
+
+        self.UNET_TARGET_REPLACE_MODULE = [
+            "Transformer2DModel",
+            "Attention",
+            "ResnetBlock2D",
+            "Downsample2D",
+            "Upsample2D",
+            "SpatialTransformer",
+        ]
+        self.TEXT_ENCODER_TARGET_REPLACE_MODULE = [
+            "ResidualAttentionBlock",
+            "CLIPAttention",
+            "CLIPMLP",
+        ]
+        self.LORA_PREFIX_UNET = "lora_unet"
+        self.LORA_PREFIX_TEXT_ENCODER = "lora_te"
+
+        def find_modules(
+            prefix, root_module: torch.nn.Module, target_replace_modules
+        ) -> dict[str, torch.nn.Module]:
+            mapping = {}
+            for name, module in root_module.named_modules():
+                if module.__class__.__name__ in target_replace_modules:
+                    for child_name, child_module in module.named_modules():
+                        layer_type = child_module.__class__.__name__
+                        if layer_type == "Linear" or (
+                            layer_type == "Conv2d"
+                            and child_module.kernel_size in [(1, 1), (3, 3)]
+                        ):
+                            lora_name = prefix + "." + name + "." + child_name
+                            lora_name = lora_name.replace(".", "_")
+                            mapping[lora_name] = child_module
+                            self.apply_module_forward(child_module, lora_name)
+            return mapping
+
+        if self.text_modules is None:
+            self.text_modules = find_modules(
+                self.LORA_PREFIX_TEXT_ENCODER,
+                text_encoder,
+                self.TEXT_ENCODER_TARGET_REPLACE_MODULE,
+            )
+
+        if self.unet_modules is None:
+            self.unet_modules = find_modules(
+                self.LORA_PREFIX_UNET, unet, self.UNET_TARGET_REPLACE_MODULE
+            )
+
+    def lora_forward_hook(self, name):
+        wrapper = self
+
+        def lora_forward(module, input_h, output):
+            if len(wrapper.loaded_loras) == 0:
+                return output
+
+            for lora in wrapper.applied_loras.values():
+                layer = lora.layers.get(name, None)
+                if layer is None:
+                    continue
+                output += layer.forward(lora, input_h)
+            return output
+
+        return lora_forward
+
+    def apply_module_forward(self, module, name):
+        handle = module.register_forward_hook(self.lora_forward_hook(name))
+        self.hooks.append(handle)
+
+    def clear_hooks(self):
+        for hook in self.hooks:
+            hook.remove()
+
+        self.hooks.clear()
+
+    def clear_applied_loras(self):
+        self.applied_loras.clear()
+
+    def clear_loaded_loras(self):
+        self.loaded_loras.clear()
+
+
+class LoRA:
+    name: str
+    layers: dict[str, LoRALayer]
+    device: torch.device
+    dtype: torch.dtype
+    wrapper: LoRAModuleWrapper
+    multiplier: float
+
+    def __init__(self, name: str, device, dtype, wrapper, multiplier=1.0):
+        self.name = name
+        self.layers = {}
+        self.multiplier = multiplier
+        self.device = device
+        self.dtype = dtype
+        self.wrapper = wrapper
+
+    def load_from_dict(self, state_dict):
+        state_dict_groupped = dict()
+
+        for key, value in state_dict.items():
+            stem, leaf = key.split(".", 1)
+            if stem not in state_dict_groupped:
+                state_dict_groupped[stem] = dict()
+            state_dict_groupped[stem][leaf] = value
+
+        for stem, values in state_dict_groupped.items():
+            if stem.startswith(self.wrapper.LORA_PREFIX_TEXT_ENCODER):
+                wrapped = self.wrapper.text_modules.get(stem, None)
+            elif stem.startswith(self.wrapper.LORA_PREFIX_UNET):
+                wrapped = self.wrapper.unet_modules.get(stem, None)
+            else:
+                continue
+
+            if wrapped is None:
+                print(f">> Missing layer: {stem}")
+                continue
+
+            # TODO: diff key
+
+            bias = None
+            alpha = None
+
+            if "alpha" in values:
+                alpha = values["alpha"].item()
+
+            if (
+                "bias_indices" in values
+                and "bias_values" in values
+                and "bias_size" in values
+            ):
+                bias = torch.sparse_coo_tensor(
+                    values["bias_indices"],
+                    values["bias_values"],
+                    tuple(values["bias_size"]),
+                ).to(device=self.device, dtype=self.dtype)
+
+            # lora and locon
+            if "lora_down.weight" in values:
+                value_down = values["lora_down.weight"]
+                value_mid = values.get("lora_mid.weight", None)
+                value_up = values["lora_up.weight"]
+
+                if type(wrapped) == torch.nn.Conv2d:
+                    if value_mid is not None:
+                        layer_down = torch.nn.Conv2d(
+                            value_down.shape[1], value_down.shape[0], (1, 1), bias=False
+                        )
+                        layer_mid = torch.nn.Conv2d(
+                            value_mid.shape[1],
+                            value_mid.shape[0],
+                            wrapped.kernel_size,
+                            wrapped.stride,
+                            wrapped.padding,
+                            bias=False,
+                        )
+                    else:
+                        layer_down = torch.nn.Conv2d(
+                            value_down.shape[1],
+                            value_down.shape[0],
+                            wrapped.kernel_size,
+                            wrapped.stride,
+                            wrapped.padding,
+                            bias=False,
+                        )
+                        layer_mid = None
+
+                    layer_up = torch.nn.Conv2d(
+                        value_up.shape[1], value_up.shape[0], (1, 1), bias=False
+                    )
+
+                elif type(wrapped) == torch.nn.Linear:
+                    layer_down = torch.nn.Linear(
+                        value_down.shape[1], value_down.shape[0], bias=False
+                    )
+                    layer_mid = None
+                    layer_up = torch.nn.Linear(
+                        value_up.shape[1], value_up.shape[0], bias=False
+                    )
+
+                else:
+                    print(
+                        f">> Encountered unknown lora layer module in {self.name}: {stem} - {type(wrapped).__name__}"
+                    )
+                    return
+
+                with torch.no_grad():
+                    layer_down.weight.copy_(value_down)
+                    if layer_mid is not None:
+                        layer_mid.weight.copy_(value_mid)
+                    layer_up.weight.copy_(value_up)
+
+                layer_down.to(device=self.device, dtype=self.dtype)
+                if layer_mid is not None:
+                    layer_mid.to(device=self.device, dtype=self.dtype)
+                layer_up.to(device=self.device, dtype=self.dtype)
+
+                rank = value_down.shape[0]
+
+                layer = LoRALayer(self.name, stem, rank, alpha)
+                # layer.bias = bias # TODO: find and debug lora/locon with bias
+                layer.down = layer_down
+                layer.mid = layer_mid
+                layer.up = layer_up
+
+            # loha
+            elif "hada_w1_b" in values:
+                rank = values["hada_w1_b"].shape[0]
+
+                layer = LoHALayer(self.name, stem, rank, alpha)
+                layer.org_module = wrapped
+                layer.bias = bias
+
+                layer.w1_a = values["hada_w1_a"].to(
+                    device=self.device, dtype=self.dtype
+                )
+                layer.w1_b = values["hada_w1_b"].to(
+                    device=self.device, dtype=self.dtype
+                )
+                layer.w2_a = values["hada_w2_a"].to(
+                    device=self.device, dtype=self.dtype
+                )
+                layer.w2_b = values["hada_w2_b"].to(
+                    device=self.device, dtype=self.dtype
+                )
+
+                if "hada_t1" in values:
+                    layer.t1 = values["hada_t1"].to(
+                        device=self.device, dtype=self.dtype
+                    )
+                else:
+                    layer.t1 = None
+
+                if "hada_t2" in values:
+                    layer.t2 = values["hada_t2"].to(
+                        device=self.device, dtype=self.dtype
+                    )
+                else:
+                    layer.t2 = None
+
+            # lokr
+            elif "lokr_w1_b" in values or "lokr_w1" in values:
+
+                if "lokr_w1_b" in values:
+                    rank = values["lokr_w1_b"].shape[0]
+                elif "lokr_w2_b" in values:
+                    rank = values["lokr_w2_b"].shape[0]
+                else:
+                    rank = None # unscaled
+
+                layer = LoKRLayer(self.name, stem, rank, alpha)
+                layer.org_module = wrapped
+                layer.bias = bias
+
+                if "lokr_w1" in values:
+                    layer.w1 = values["lokr_w1"].to(device=self.device, dtype=self.dtype)
+                else:
+                    layer.w1_a = values["lokr_w1_a"].to(device=self.device, dtype=self.dtype)
+                    layer.w1_b = values["lokr_w1_b"].to(device=self.device, dtype=self.dtype)
+
+                if "lokr_w2" in values:
+                    layer.w2 = values["lokr_w2"].to(device=self.device, dtype=self.dtype)
+                else:
+                    layer.w2_a = values["lokr_w2_a"].to(device=self.device, dtype=self.dtype)
+                    layer.w2_b = values["lokr_w2_b"].to(device=self.device, dtype=self.dtype)
+
+                if "lokr_t2" in values:
+                    layer.t2 = values["lokr_t2"].to(device=self.device, dtype=self.dtype)
+
+
+            else:
+                print(
+                    f">> Encountered unknown lora layer module in {self.name}: {stem} - {type(wrapped).__name__}"
+                )
+                return
+
+            self.layers[stem] = layer
+
+
+class KohyaLoraManager:
+    lora_path = Path(global_lora_models_dir())
+    vector_length_cache_path = lora_path / '.vectorlength.cache'
+    
+    def __init__(self, pipe):
+        self.unet = pipe.unet
+        self.wrapper = LoRAModuleWrapper(pipe.unet, pipe.text_encoder)
+        self.text_encoder = pipe.text_encoder
+        self.device = torch.device(choose_torch_device())
+        self.dtype = pipe.unet.dtype
+
+    def load_lora_module(self, name, path_file, multiplier: float = 1.0):
+        print(f"   | Found lora {name} at {path_file}")
+        if path_file.suffix == ".safetensors":
+            checkpoint = load_file(path_file.absolute().as_posix(), device="cpu")
+        else:
+            checkpoint = torch.load(path_file, map_location="cpu")
+
+        if not self.check_model_compatibility(checkpoint):
+            raise IncompatibleModelException
+
+        lora = LoRA(name, self.device, self.dtype, self.wrapper, multiplier)
+        lora.load_from_dict(checkpoint)
+        self.wrapper.loaded_loras[name] = lora
+
+        return lora
+
+    def apply_lora_model(self, name, mult: float = 1.0):
+        path_file = None
+        for suffix in ["ckpt", "safetensors", "pt"]:
+            path_files = [x for x in Path(self.lora_path).glob(f"**/{name}.{suffix}")]
+            if len(path_files):
+                path_file = path_files[0]
+                print(f"   | Loading lora {path_file.name} with weight {mult}")
+                break
+        if not path_file:
+            print(f"   ** Unable to find lora: {name}")
+            return
+
+        lora = self.wrapper.loaded_loras.get(name, None)
+        if lora is None:
+            lora = self.load_lora_module(name, path_file, mult)
+
+        lora.multiplier = mult
+        self.wrapper.applied_loras[name] = lora
+
+    def unload_applied_lora(self, lora_name: str) -> bool:
+        """If the indicated LoRA has previously been applied then
+        unload it and return True. Return False if the LoRA was
+        not previously applied (for status reporting)
+        """
+        if lora_name in self.wrapper.applied_loras:
+            del self.wrapper.applied_loras[lora_name]
+            return True
+        return False
+
+    def unload_lora(self, lora_name: str) -> bool:
+        if lora_name in self.wrapper.loaded_loras:
+            del self.wrapper.loaded_loras[lora_name]
+            return True
+        return False
+
+    def clear_loras(self):
+        self.wrapper.clear_applied_loras()
+
+    def check_model_compatibility(self, checkpoint) -> bool:
+        """Checks whether the LoRA checkpoint is compatible with the token vector
+        length of the model that this manager is associated with.
+        """
+        model_token_vector_length = (
+            self.text_encoder.get_input_embeddings().weight.data[0].shape[0]
+        )
+        lora_token_vector_length = self.vector_length_from_checkpoint(checkpoint)
+        return model_token_vector_length == lora_token_vector_length
+
+    @staticmethod
+    def vector_length_from_checkpoint(checkpoint: dict) -> int:
+        """Return the vector token length for the passed LoRA checkpoint object.
+        This is used to determine which SD model version the LoRA was based on.
+        768 -> SDv1
+        1024-> SDv2
+        """
+        key1 = "lora_te_text_model_encoder_layers_0_mlp_fc1.lora_down.weight"
+        key2 = "lora_te_text_model_encoder_layers_0_self_attn_k_proj.hada_w1_a"
+        lora_token_vector_length = (
+            checkpoint[key1].shape[1]
+            if key1 in checkpoint
+            else checkpoint[key2].shape[0]
+            if key2 in checkpoint
+            else 768
+        )
+        return lora_token_vector_length
+
+    @classmethod
+    def vector_length_from_checkpoint_file(self, checkpoint_path: Path) -> int:
+        with LoraVectorLengthCache(self.vector_length_cache_path) as cache:
+            if str(checkpoint_path) not in cache:
+                if checkpoint_path.suffix == ".safetensors":
+                    checkpoint = load_file(
+                        checkpoint_path.absolute().as_posix(), device="cpu"
+                    )
+                else:
+                    checkpoint = torch.load(checkpoint_path, map_location="cpu")
+                cache[str(checkpoint_path)] = KohyaLoraManager.vector_length_from_checkpoint(
+                    checkpoint
+                )
+            return cache[str(checkpoint_path)]
+
+class LoraVectorLengthCache(object):
+    def __init__(self, cache_path: Path):
+        self.cache_path = cache_path
+        self.lock = FileLock(Path(cache_path.parent, ".cachelock"))
+        self.cache = {}
+
+    def __enter__(self):
+        self.lock.acquire(timeout=10)
+        try:
+            if self.cache_path.exists():
+                with open(self.cache_path, "r") as json_file:
+                    self.cache = json.load(json_file)
+        except Timeout:
+            print(
+                "** Can't acquire lock on lora vector length cache. Operations will be slower"
+            )
+        except (json.JSONDecodeError, OSError):
+            self.cache_path.unlink()
+        return self.cache
+
+    def __exit__(self, type, value, traceback):
+        with open(self.cache_path, "w") as json_file:
+            json.dump(self.cache, json_file)
+        self.lock.release()

ldm/modules/lora_manager.py

@@ -0,0 +1,101 @@
+import os
+from diffusers import StableDiffusionPipeline
+from pathlib import Path
+
+from diffusers import UNet2DConditionModel, StableDiffusionPipeline
+from ldm.invoke.globals import global_lora_models_dir
+from .kohya_lora_manager import KohyaLoraManager, IncompatibleModelException
+from typing import Optional, Dict
+
+class LoraCondition:
+    name: str
+    weight: float
+
+    def __init__(self,
+                 name,
+                 weight: float = 1.0,
+                 unet: UNet2DConditionModel=None,  # for diffusers format LoRAs
+                 kohya_manager: Optional[KohyaLoraManager]=None,  # for KohyaLoraManager-compatible LoRAs
+                 ):
+        self.name = name
+        self.weight = weight
+        self.kohya_manager = kohya_manager
+        self.unet = unet
+
+    def __call__(self):
+        # TODO: make model able to load from huggingface, rather then just local files
+        path = Path(global_lora_models_dir(), self.name)
+        if path.is_dir():
+            if not self.unet:
+                print(f"   ** Unable to load diffusers-format LoRA {self.name}: unet is None")
+                return
+            if self.unet.load_attn_procs:
+                file = Path(path, "pytorch_lora_weights.bin")
+                if file.is_file():
+                    print(f">> Loading LoRA: {path}")
+                    self.unet.load_attn_procs(path.absolute().as_posix())
+                else:
+                    print(f"   ** Unable to find valid LoRA at: {path}")
+            else:
+                print("   ** Invalid Model to load LoRA")
+        elif self.kohya_manager:
+            try:
+                self.kohya_manager.apply_lora_model(self.name,self.weight)
+            except IncompatibleModelException:
+                print(f"   ** LoRA {self.name} is incompatible with this model; will generate without the LoRA applied.")
+        else:
+            print("   ** Unable to load LoRA")
+
+    def unload(self):
+        if self.kohya_manager and self.kohya_manager.unload_applied_lora(self.name):
+            print(f'>> unloading LoRA {self.name}')
+            
+
+class LoraManager:
+    def __init__(self, pipe: StableDiffusionPipeline):
+        # Kohya class handles lora not generated through diffusers
+        self.kohya = KohyaLoraManager(pipe)
+        self.unet = pipe.unet
+
+    def set_loras_conditions(self, lora_weights: list):
+        conditions = []
+        if len(lora_weights) > 0:
+            for lora in lora_weights:
+                conditions.append(LoraCondition(lora.model, lora.weight, self.unet, self.kohya))
+
+        if len(conditions) > 0:
+            return conditions
+
+        return None
+    
+    def list_compatible_loras(self)->Dict[str, Path]:
+        '''
+        List all the LoRAs in the global lora directory that
+        are compatible with the current model. Return a dictionary
+        of the lora basename and its path.
+        '''
+        model_length = self.kohya.text_encoder.get_input_embeddings().weight.data[0].shape[0]
+        return self.list_loras(model_length)
+
+    @staticmethod
+    def list_loras(token_vector_length:int=None)->Dict[str, Path]:
+        '''List the LoRAS in the global lora directory.
+        If token_vector_length is provided, then only return
+        LoRAS that have the indicated length:
+        768: v1 models
+        1024: v2 models
+        '''
+        path = Path(global_lora_models_dir())
+        models_found = dict()
+        for root,_,files in os.walk(path):
+            for x in files:
+                name = Path(x).stem
+                suffix = Path(x).suffix
+                if suffix not in [".ckpt", ".pt", ".safetensors"]:
+                    continue
+                path = Path(root,x)
+                if token_vector_length is None:
+                    models_found[name]=Path(root,x)  # unconditional addition
+                elif token_vector_length == KohyaLoraManager.vector_length_from_checkpoint_file(path):
+                    models_found[name]=Path(root,x)  # conditional on the base model matching
+        return models_found

ldm/modules/peft_manager.py

@@ -0,0 +1,95 @@
+from peft import LoraModel, LoraConfig, set_peft_model_state_dict
+import torch
+import json
+from pathlib import Path
+from ldm.invoke.globals import global_lora_models_dir
+
+class LoraPeftModule:
+    def __init__(self, lora_dir, multiplier: float = 1.0):
+        self.lora_dir = lora_dir
+        self.multiplier = multiplier
+        self.config = self.load_config()
+        self.checkpoint = self.load_checkpoint()
+
+    def load_config(self):
+        lora_config_file = Path(self.lora_dir, f'lora_config.json')
+        with open(lora_config_file, "r") as f:
+            return json.load(f)
+
+    def load_checkpoint(self):
+        return torch.load(Path(self.lora_dir, f'lora.pt'))
+
+    def unet(self, text_encoder):
+        lora_ds = {
+            k.replace("text_encoder_", ""): v for k, v in self.checkpoint.items() if "text_encoder_" in k
+        }
+        config = LoraConfig(**self.config["peft_config"])
+        model = LoraModel(config, text_encoder)
+        set_peft_model_state_dict(model, lora_ds)
+        return model
+
+    def text_encoder(self, unet):
+        lora_ds = {
+            k: v for k, v in self.checkpoint.items() if "text_encoder_" not in k
+        }
+        config = LoraConfig(**self.config["text_encoder_peft_config"])
+        model = LoraModel(config, unet)
+        set_peft_model_state_dict(model, lora_ds)
+        return model
+
+    def apply(self, pipe, dtype):
+        pipe.unet = self.unet(pipe.unet)
+        if "text_encoder_peft_config" in self.config:
+            pipe.text_encoder = self.text_encoder(pipe.text_encoder)
+
+        if dtype in (torch.float16, torch.bfloat16):
+            pipe.unet.half()
+            pipe.text_encoder.half()
+
+        return pipe
+
+
+class PeftManager:
+    modules: list[LoraPeftModule]
+
+    def __init__(self):
+        self.lora_path = global_lora_models_dir()
+        self.modules = []
+
+    def set_loras(self, lora_weights: list):
+        if len(lora_weights) > 0:
+            for lora in lora_weights:
+                self.add(lora.model, lora.weight)
+
+    def add(self, name, multiplier: float = 1.0):
+        lora_dir = Path(self.lora_path, name)
+
+        if lora_dir.exists():
+            lora_config_file = Path(lora_dir, f'lora_config.json')
+            lora_checkpoint = Path(lora_dir, f'lora.pt')
+
+            if lora_config_file.exists() and lora_checkpoint.exists():
+                self.modules.append(LoraPeftModule(lora_dir, multiplier))
+                return
+
+        print(f">> Failed to load lora {name}")
+
+    def load(self, pipe, dtype):
+        if len(self.modules) > 0:
+            for module in self.modules:
+                pipe = module.apply(pipe, dtype)
+
+        return pipe
+
+    # Simple check to allow previous functionality
+    def should_use(self, lora_weights: list):
+        if len(lora_weights) > 0:
+            for lora in lora_weights:
+                lora_dir = Path(self.lora_path, lora.model)
+                if lora_dir.exists():
+                    lora_config_file = Path(lora_dir, f'lora_config.json')
+                    lora_checkpoint = Path(lora_dir, f'lora.pt')
+                    if lora_config_file.exists() and lora_checkpoint.exists():
+                        return False
+
+        return True

ldm/modules/textual_inversion_manager.py

@@ -1,9 +1,9 @@
-import os
 import traceback
 from dataclasses import dataclass
 from pathlib import Path
 from typing import Optional, Union
 
+import safetensors.torch
 import torch
 from picklescan.scanner import scan_file_path
 from transformers import CLIPTextModel, CLIPTokenizer
@@ -71,21 +71,6 @@ class TextualInversionManager(BaseTextualInversionManager):
 
         if str(ckpt_path).endswith(".DS_Store"):
             return
-
-        try:
-            scan_result = scan_file_path(str(ckpt_path))
-            if scan_result.infected_files == 1:
-                print(
-                    f"\n### Security Issues Found in Model: {scan_result.issues_count}"
-                )
-                print("### For your safety, InvokeAI will not load this embed.")
-                return
-        except Exception:
-            print(
-                f"### {ckpt_path.parents[0].name}/{ckpt_path.name} is damaged or corrupt."
-            )
-            return
-
         embedding_info = self._parse_embedding(str(ckpt_path))
 
         if embedding_info is None:
@@ -96,7 +81,7 @@ class TextualInversionManager(BaseTextualInversionManager):
             != embedding_info["token_dim"]
         ):
             print(
-                f"** Notice: {ckpt_path.parents[0].name}/{ckpt_path.name} was trained on a model with an incompatible token dimension: {self.text_encoder.get_input_embeddings().weight.data[0].shape[0]} vs {embedding_info['token_dim']}."
+                f"   ** Notice: {ckpt_path.parents[0].name}/{ckpt_path.name} was trained on a model with an incompatible token dimension: {self.text_encoder.get_input_embeddings().weight.data[0].shape[0]} vs {embedding_info['token_dim']}."
             )
             return
 
@@ -309,92 +294,72 @@ class TextualInversionManager(BaseTextualInversionManager):
 
         return token_id
 
-    def _parse_embedding(self, embedding_file: str):
-        file_type = embedding_file.split(".")[-1]
-        if file_type == "pt":
-            return self._parse_embedding_pt(embedding_file)
-        elif file_type == "bin":
-            return self._parse_embedding_bin(embedding_file)
-        else:
-            print(f"** Notice: unrecognized embedding file format: {embedding_file}")
+    def _parse_embedding(self, embedding_file: str)->dict:
+        suffix = Path(embedding_file).suffix
+        try:
+            if suffix in [".pt",".ckpt",".bin"]:
+                scan_result = scan_file_path(embedding_file)
+                if scan_result.infected_files == 1:
+                    print(
+                        f"   ** Security Issues Found in Model: {scan_result.issues_count}"
+                    )
+                    print("   ** For your safety, InvokeAI will not load this embed.")
+                    return
+                ckpt = torch.load(embedding_file,map_location="cpu")
+            else:
+                ckpt = safetensors.torch.load_file(embedding_file)
+        except Exception as e:
+            print(f"   ** Notice: unrecognized embedding file format: {embedding_file}: {e}")
             return None
-
-    def _parse_embedding_pt(self, embedding_file):
-        embedding_ckpt = torch.load(embedding_file, map_location="cpu")
-        embedding_info = {}
-
-        # Check if valid embedding file
-        if "string_to_token" and "string_to_param" in embedding_ckpt:
-            # Catch variants that do not have the expected keys or values.
-            try:
-                embedding_info["name"] = embedding_ckpt["name"] or os.path.basename(
-                    os.path.splitext(embedding_file)[0]
-                )
-
-                # Check num of embeddings and warn user only the first will be used
-                embedding_info["num_of_embeddings"] = len(
-                    embedding_ckpt["string_to_token"]
-                )
-                if embedding_info["num_of_embeddings"] > 1:
-                    print(">> More than 1 embedding found. Will use the first one")
-
-                embedding = list(embedding_ckpt["string_to_param"].values())[0]
-            except (AttributeError, KeyError):
-                return self._handle_broken_pt_variants(embedding_ckpt, embedding_file)
-
-            embedding_info["embedding"] = embedding
-            embedding_info["num_vectors_per_token"] = embedding.size()[0]
-            embedding_info["token_dim"] = embedding.size()[1]
-
-            try:
-                embedding_info["trained_steps"] = embedding_ckpt["step"]
-                embedding_info["trained_model_name"] = embedding_ckpt[
-                    "sd_checkpoint_name"
-                ]
-                embedding_info["trained_model_checksum"] = embedding_ckpt[
-                    "sd_checkpoint"
-                ]
-            except AttributeError:
-                print(">> No Training Details Found. Passing ...")
-
-        # .pt files found at https://cyberes.github.io/stable-diffusion-textual-inversion-models/
-        # They are actually .bin files
-        elif len(embedding_ckpt.keys()) == 1:
-            embedding_info = self._parse_embedding_bin(embedding_file)
-
+        
+        # try to figure out what kind of embedding file it is and parse accordingly
+        keys = list(ckpt.keys())
+        if all(x in keys for x in ['string_to_token','string_to_param','name','step']):
+            return self._parse_embedding_v1(ckpt, embedding_file)     # example rem_rezero.pt
+        
+        elif all(x in keys for x in ['string_to_token','string_to_param']):
+            return self._parse_embedding_v2(ckpt, embedding_file)     # example midj-strong.pt
+        
+        elif 'emb_params' in keys:
+            return self._parse_embedding_v3(ckpt, embedding_file)     # example easynegative.safetensors
+        
         else:
-            print(">> Invalid embedding format")
-            embedding_info = None
+            return self._parse_embedding_v4(ckpt, embedding_file)     # usually a '.bin' file
 
+    def _parse_embedding_v1(self, embedding_ckpt: dict, file_path: str):
+        basename = Path(file_path).stem
+        print(f'   | Loading v1 embedding file: {basename}')
+
+        embedding_info = {}
+        embedding_info["name"] = embedding_ckpt["name"]
+                
+        # Check num of embeddings and warn user only the first will be used
+        embedding_info["num_of_embeddings"] = len(
+            embedding_ckpt["string_to_token"]
+        )
+        if embedding_info["num_of_embeddings"] > 1:
+            print("   | More than 1 embedding found. Will use the first one")
+        embedding = list(embedding_ckpt["string_to_param"].values())[0]
+        embedding_info["embedding"] = embedding
+        embedding_info["num_vectors_per_token"] = embedding.size()[0]
+        embedding_info["token_dim"] = embedding.size()[1]
+        embedding_info["trained_steps"] = embedding_ckpt["step"]
+        embedding_info["trained_model_name"] = embedding_ckpt[
+            "sd_checkpoint_name"
+        ]
+        embedding_info["trained_model_checksum"] = embedding_ckpt[
+            "sd_checkpoint"
+        ]
         return embedding_info
 
-    def _parse_embedding_bin(self, embedding_file):
-        embedding_ckpt = torch.load(embedding_file, map_location="cpu")
-        embedding_info = {}
-
-        if list(embedding_ckpt.keys()) == 0:
-            print(">> Invalid concepts file")
-            embedding_info = None
-        else:
-            for token in list(embedding_ckpt.keys()):
-                embedding_info["name"] = (
-                    token
-                    or f"<{os.path.basename(os.path.splitext(embedding_file)[0])}>"
-                )
-                embedding_info["embedding"] = embedding_ckpt[token]
-                embedding_info[
-                    "num_vectors_per_token"
-                ] = 1  # All Concepts seem to default to 1
-                embedding_info["token_dim"] = embedding_info["embedding"].size()[0]
-
-        return embedding_info
-
-    def _handle_broken_pt_variants(
-        self, embedding_ckpt: dict, embedding_file: str
+    def _parse_embedding_v2 (
+        self, embedding_ckpt: dict, file_path: str
     ) -> dict:
         """
-        This handles the broken .pt file variants. We only know of one at present.
+        This handles embedding .pt file variant #2.
         """
+        basename = Path(file_path).stem
+        print(f'   | Loading v2 embedding file: {basename}')
         embedding_info = {}
         if isinstance(
             list(embedding_ckpt["string_to_token"].values())[0], torch.Tensor
@@ -403,7 +368,7 @@ class TextualInversionManager(BaseTextualInversionManager):
                 embedding_info["name"] = (
                     token
                     if token != "*"
-                    else f"<{os.path.basename(os.path.splitext(embedding_file)[0])}>"
+                    else f"<{basename}>"
                 )
                 embedding_info["embedding"] = embedding_ckpt[
                     "string_to_param"
@@ -413,7 +378,46 @@ class TextualInversionManager(BaseTextualInversionManager):
                 ].shape[0]
                 embedding_info["token_dim"] = embedding_info["embedding"].size()[1]
         else:
-            print(">> Invalid embedding format")
+            print(f"   ** {basename}: Unrecognized embedding format")
             embedding_info = None
 
         return embedding_info
+
+    def _parse_embedding_v3(self, embedding_ckpt: dict, file_path: str):
+        """
+        Parse 'version 3' of the .pt textual inversion embedding files.
+        """
+        basename = Path(file_path).stem
+        print(f'   | Loading v3 embedding file: {basename}')
+        embedding_info = {}
+        embedding_info["name"] = f'<{basename}>'
+        embedding_info["num_of_embeddings"] = 1
+        embedding = embedding_ckpt['emb_params']
+        embedding_info["embedding"] = embedding
+        embedding_info["num_vectors_per_token"] = embedding.size()[0]
+        embedding_info["token_dim"] = embedding.size()[1]
+        return embedding_info
+    
+    def _parse_embedding_v4(self, embedding_ckpt: dict, filepath: str):
+        """
+        Parse 'version 4' of the textual inversion embedding files. This one
+        is usually associated with .bin files trained by HuggingFace diffusers.
+        """
+        basename = Path(filepath).stem
+        short_path = Path(filepath).parents[0].name+'/'+Path(filepath).name
+        
+        print(f'   | Loading v4 embedding file: {short_path}')
+        embedding_info = {}
+        if list(embedding_ckpt.keys()) == 0:
+            print(f"   ** Invalid embeddings file: {short_path}")
+            embedding_info = None
+        else:
+            for token in list(embedding_ckpt.keys()):
+                embedding_info["name"] = (
+                    token
+                    or f"<{basename}>"
+                )
+                embedding_info["embedding"] = embedding_ckpt[token]
+                embedding_info["num_vectors_per_token"] = 1  # All Concepts seem to default to 1
+                embedding_info["token_dim"] = embedding_info["embedding"].size()[0]
+        return embedding_info

pyproject.toml

@@ -32,13 +32,16 @@ dependencies = [
   "albumentations",
   "click",
   "clip_anytorch",
-  "compel==0.1.7",
+  "compel~=1.1.0",
   "datasets",
-  "diffusers[torch]~=0.14",
+  "diffusers[torch]==0.14",
   "dnspython==2.2.1",
   "einops",
   "eventlet",
   "facexlib",
+  "fastapi==0.85.0",
+  "fastapi-events==0.6.0",
+  "fastapi-socketio==0.0.9",
   "flask==2.1.3",
   "flask_cors==3.0.10",
   "flask_socketio==5.3.0",
@@ -54,6 +57,7 @@ dependencies = [
   "numpy<1.24",
   "omegaconf",
   "opencv-python",
+  "peft",
   "picklescan",
   "pillow",
   "prompt-toolkit",
@@ -61,6 +65,7 @@ dependencies = [
   "packaging",
   "pypatchmatch",
   "pyreadline3",
+  "python-multipart==0.0.5",
   "pytorch-lightning==1.7.7",
   "realesrgan",
   "requests==2.28.2",
@@ -122,6 +127,7 @@ requires-python = ">=3.9, <3.11"
 "invokeai-ti" = "ldm.invoke.training.textual_inversion:main"
 "invokeai-update" = "ldm.invoke.config.invokeai_update:main"
 "invokeai-batch" = "ldm.invoke.dynamic_prompts:main"
+"invokeai-metadata" = "ldm.invoke.invokeai_metadata:main"
 
 [project.urls]
 "Bug Reports" = "https://github.com/invoke-ai/InvokeAI/issues"

scripts/invoke-new.py

@@ -0,0 +1,20 @@
+# Copyright (c) 2022 Kyle Schouviller (https://github.com/kyle0654)
+
+import os
+import sys
+
+def main():
+    # Change working directory to the repo root
+    os.chdir(os.path.abspath(os.path.join(os.path.dirname(__file__), "..")))
+
+    if '--web' in sys.argv:
+        from ldm.invoke.app.api_app import invoke_api
+        invoke_api()
+    else:
+        # TODO: Parse some top-level args here.
+        from ldm.invoke.app.cli_app import invoke_cli
+        invoke_cli()
+
+
+if __name__ == '__main__':
+    main()

static/dream_web/test.html

@@ -0,0 +1,206 @@
+<html lang="en">
+
+<head>
+  <title>InvokeAI Test</title>
+  <meta charset="utf-8">
+  <link rel="icon" type="image/x-icon" href="static/dream_web/favicon.ico" />
+  <meta name="viewport" content="width=device-width, initial-scale=1.0">
+
+  <!--<script src="config.js"></script>-->
+  <script src="https://cdnjs.cloudflare.com/ajax/libs/socket.io/4.0.1/socket.io.js"
+    integrity="sha512-q/dWJ3kcmjBLU4Qc47E4A9kTB4m3wuTY7vkFJDTZKjTs8jhyGQnaUrxa0Ytd0ssMZhbNua9hE+E7Qv1j+DyZwA=="
+    crossorigin="anonymous"></script>
+
+  <script src="https://cdnjs.cloudflare.com/ajax/libs/fslightbox/3.0.9/index.js"></script>
+
+  <style>
+      body {
+          background:#334;
+      }
+
+      #gallery > a.image {
+          display:inline-block;
+          width:128px;
+          height:128px;
+          margin:8px;
+          background-size:contain;
+          background-repeat:no-repeat;
+          background-position:center;
+      }
+
+      .results {
+        border-color:green;
+        border-width:1px;
+        border-style:solid;
+      }
+
+      .intermediates {
+        border-color:yellow;
+        border-width:1px;
+        border-style:solid;
+      }
+  </style>
+</head>
+
+<body>
+    <button onclick="dostuff();">Test Invoke</button>
+    <div id="gallery"></div>
+    <div id="textlog">
+        <textarea id='log' rows=10 cols=60 autofocus>
+        </textarea>
+    </div>
+
+    <script type="text/javascript">
+        const socket_url = `ws://${window.location.host}`;
+        const socket = io(socket_url, {
+            path: "/ws/socket.io"
+        });
+        socket.on('connect', (data) => {
+            //socket.emit('subscribe', { 'session': 'WcBtYATwT92Mrb9zLgeyNw==' });
+        });
+
+        function addGalleryImage(src, className) {
+                let gallery = document.getElementById("gallery");
+                let div = document.createElement("a");
+                div.href = src;
+                div.setAttribute('data-fslightbox', '');
+                div.classList.add("image");
+                div.classList.add(className);
+                div.style.backgroundImage = `url('${src}')`;
+                gallery.appendChild(div);
+
+                refreshFsLightbox();
+        }
+
+        function log(event, data) {
+            document.getElementById("log").value += `${event} => ${JSON.stringify(data)}\n`;
+
+            if (data.result?.image?.image_type) {
+                addGalleryImage(`/api/v1/images/${data.result.image.image_type}/${data.result.image.image_name}`, data.result.image.image_type);
+            }
+            if (data.result?.mask?.image_type) {
+                addGalleryImage(`/api/v1/images/${data.result.mask.image_type}/${data.result.mask.image_name}`, data.result.mask.image_type);
+            }
+
+            console.log(`${event} => ${JSON.stringify(data)}`);
+        }
+
+        socket.on('generator_progress', (data) => log('generator_progress', data));
+        socket.on('invocation_complete', (data) => log('invocation_complete', data));
+        socket.on('invocation_started', (data) => log('invocation_started', data));
+        socket.on('session_complete', (data) => {
+            log('session_complete', data);
+
+            // NOTE: you may not want to unsubscribe if you plan to continue using this session,
+            //       just make sure you unsubscribe eventually
+            socket.emit('unsubscribe', { 'session': data.session_id });
+        });
+
+        function dostuff() {
+            let prompt = "hyper detailed 4k cryengine 3D render of a cat in a dune buggy, trending on artstation, soft atmospheric lighting, volumetric lighting, cinematic still, golden hour, crepuscular rays, smooth [grainy]";
+            let strength = 0.95;
+            let sampler = 'keuler_a';
+            let steps = 50;
+            let seed = -1;
+
+            // Start building nodes
+            var id = 1;
+            var initialNode = {"id": id.toString(), "type": "txt2img", "prompt": prompt, "sampler": sampler, "steps": steps, "seed": seed};
+            id++;
+            var upscaleNode = {"id": id.toString(), "type": "show_image" };
+            id++
+
+            nodes = {};
+            nodes[initialNode.id] = initialNode;
+            nodes[upscaleNode.id] = upscaleNode;
+            links = [
+                [{ "node_id": initialNode.id, field: "image" },
+                 { "node_id": upscaleNode.id, field: "image" }]
+            ];
+            // expandSize = 128;
+            // for (var i = 0; i < 6; ++i) {
+            //     var i_seed = (seed == -1) ? -1 : seed + i + 1;
+            //     var startid = id - 1;
+            //     var offset = (i < 3) ? -expandSize : (3 * expandSize) + ((i - 3 + 1) * expandSize);
+            //     nodes.push({"id": id.toString(), "type": "crop", "x": offset, "y": 0, "width": 512, "height": 512});
+            //     let id_gen = id;
+            //     links.push({"from_node": {"id": startid.toString(), "field": "image"},"to_node": {"id": id_gen.toString(),"field": "image"}});
+            //     id++;
+                
+            //     nodes.push({"id": id.toString(), "type": "tomask"});
+            //     let id_mask = id;
+            //     links.push({"from_node": {"id": id_gen.toString(), "field": "image"},"to_node": {"id": id_mask.toString(),"field": "image"}});
+            //     id++;
+
+            //     nodes.push({"id": id.toString(), "type": "blur", "radius": 32});
+            //     let id_mask_blur = id;
+            //     links.push({"from_node": {"id": id_mask.toString(), "field": "mask"},"to_node": {"id": id_mask_blur.toString(),"field": "image"}});
+            //     id++
+
+            //     nodes.push({"id": id.toString(), "type": "ilerp", "min": 128, "max": 255});
+            //     let id_ilerp = id;
+            //     links.push({"from_node": {"id": id_mask_blur.toString(), "field": "image"},"to_node": {"id": id_ilerp.toString(),"field": "image"}});
+            //     id++
+
+            //     nodes.push({"id": id.toString(), "type": "cv_inpaint"});
+            //     let id_cv_inpaint = id;
+            //     links.push({"from_node": {"id": id_gen.toString(), "field": "image"},"to_node": {"id": id_cv_inpaint.toString(),"field": "image"}});
+            //     links.push({"from_node": {"id": id_mask.toString(), "field": "mask"},"to_node": {"id": id_cv_inpaint.toString(),"field": "mask"}});
+            //     id++;
+
+            //     nodes.push({"id": id.toString(), "type": "img2img", "prompt": prompt, "strength": strength, "sampler": sampler, "steps": steps, "seed": i_seed, "color_match": true, "inpaint_replace": inpaint_replace});
+            //     let id_i2i = id;
+            //     links.push({"from_node": {"id": id_cv_inpaint.toString(), "field": "image"},"to_node": {"id": id_i2i.toString(),"field": "image"}});
+            //     links.push({"from_node": {"id": id_ilerp.toString(), "field": "image"},"to_node": {"id": id_i2i.toString(),"field": "mask"}});
+            //     id++;
+
+            //     nodes.push({"id": id.toString(), "type": "paste", "x": offset, "y": 0});
+            //     let id_paste = id;
+            //     links.push({"from_node": {"id": startid.toString(), "field": "image"},"to_node": {"id": id_paste.toString(),"field": "base_image"}});
+            //     links.push({"from_node": {"id": id_i2i.toString(), "field": "image"},"to_node": {"id": id_paste.toString(),"field": "image"}});
+            //     links.push({"from_node": {"id": id_ilerp.toString(), "field": "image"},"to_node": {"id": id_paste.toString(),"field": "mask"}});
+            //     id++;
+            // }
+
+            var graph = {
+                "nodes": nodes,
+                "edges": links
+            };
+
+            // var defaultGraph = {"nodes": [
+            //     {"id": "1", "type": "txt2img", "prompt": prompt},
+            //     {"id": "2", "type": "crop", "x": -256, "y": 128, "width": 512, "height": 512},
+            //     {"id": "3", "type": "tomask"},
+            //     {"id": "4", "type": "cv_inpaint"},
+            //     {"id": "5", "type": "img2img", "prompt": prompt, "strength": 0.9},
+            //     {"id": "6", "type": "paste", "x": -256, "y": 128},
+            // ],
+            // "links": [
+            //     {"from_node": {"id": "1","field": "image"},"to_node": {"id": "2","field": "image"}},
+            //     {"from_node": {"id": "2","field": "image"},"to_node": {"id": "3","field": "image"}},
+            //     {"from_node": {"id": "2","field": "image"},"to_node": {"id": "4","field": "image"}},
+            //     {"from_node": {"id": "3","field": "mask"},"to_node": {"id": "4","field": "mask"}},
+            //     {"from_node": {"id": "4","field": "image"},"to_node": {"id": "5","field": "image"}},
+            //     {"from_node": {"id": "3","field": "mask"},"to_node": {"id": "5","field": "mask"}},
+            //     {"from_node": {"id": "1","field": "image"},"to_node": {"id": "6","field": "base_image"}},
+            //     {"from_node": {"id": "5","field": "image"},"to_node": {"id": "6","field": "image"}}
+            // ]};
+            fetch('/api/v1/sessions/', {
+                method: 'POST',
+                headers: new Headers({'content-type': 'application/json'}),
+                body: JSON.stringify(graph)
+            }).then(response => response.json())
+            .then(data => {
+                sessionId = data.id;
+                socket.emit('subscribe', { 'session': sessionId });
+                fetch(`/api/v1/sessions/${sessionId}/invoke?all=true`, {
+                    method: 'PUT',
+                    headers: new Headers({'content-type': 'application/json'}),
+                    body: ''
+                });
+            });
+        }
+    </script>
+</body>
+
+</html>