chore: release v4.2.9.dev12

- Rely on redux + reselect more
- Remove all nanostores that simply "mirrored" redux state in favor of direct subscriptions to redux store
- Add abstractions for creating redux subs and running selectors
- Add `initialize` method to CanvasModuleBase, for post-instantiation tasks
- Reduce local caching of state in modules to a minimum

.github/workflows/build-container.yml

@@ -13,6 +13,12 @@ on:
     tags:
       - 'v*.*.*'
   workflow_dispatch:
+    inputs:
+      push-to-registry:
+        description: Push the built image to the container registry
+        required: false
+        type: boolean
+        default: false
 
 permissions:
   contents: write
@@ -50,16 +56,15 @@ jobs:
           df -h
 
       - name: Checkout
-        uses: actions/checkout@v3
+        uses: actions/checkout@v4
 
       - name: Docker meta
         id: meta
-        uses: docker/metadata-action@v4
+        uses: docker/metadata-action@v5
         with:
           github-token: ${{ secrets.GITHUB_TOKEN }}
           images: |
             ghcr.io/${{ github.repository }}
-            ${{ env.DOCKERHUB_REPOSITORY }}
           tags: |
             type=ref,event=branch
             type=ref,event=tag
@@ -72,49 +77,33 @@ jobs:
             suffix=-${{ matrix.gpu-driver }},onlatest=false
 
       - name: Set up QEMU
-        uses: docker/setup-qemu-action@v2
+        uses: docker/setup-qemu-action@v3
 
       - name: Set up Docker Buildx
-        uses: docker/setup-buildx-action@v2
+        uses: docker/setup-buildx-action@v3
         with:
           platforms: ${{ env.PLATFORMS }}
 
       - name: Login to GitHub Container Registry
         if: github.event_name != 'pull_request'
-        uses: docker/login-action@v2
+        uses: docker/login-action@v3
         with:
           registry: ghcr.io
           username: ${{ github.repository_owner }}
           password: ${{ secrets.GITHUB_TOKEN }}
 
-      # - name: Login to Docker Hub
-      #   if: github.event_name != 'pull_request' && vars.DOCKERHUB_REPOSITORY != ''
-      #   uses: docker/login-action@v2
-      #   with:
-      #     username: ${{ secrets.DOCKERHUB_USERNAME }}
-      #     password: ${{ secrets.DOCKERHUB_TOKEN }}
-
       - name: Build container
         timeout-minutes: 40
         id: docker_build
-        uses: docker/build-push-action@v4
+        uses: docker/build-push-action@v6
         with:
           context: .
           file: docker/Dockerfile
           platforms: ${{ env.PLATFORMS }}
-          push: ${{ github.ref == 'refs/heads/main' || github.ref_type == 'tag' }}
+          push: ${{ github.ref == 'refs/heads/main' || github.ref_type == 'tag' || github.event.inputs.push-to-registry }}
           tags: ${{ steps.meta.outputs.tags }}
           labels: ${{ steps.meta.outputs.labels }}
           cache-from: |
             type=gha,scope=${{ github.ref_name }}-${{ matrix.gpu-driver }}
             type=gha,scope=main-${{ matrix.gpu-driver }}
           cache-to: type=gha,mode=max,scope=${{ github.ref_name }}-${{ matrix.gpu-driver }}
-
-      # - name: Docker Hub Description
-      #   if: github.ref == 'refs/heads/main' || github.ref == 'refs/tags/*' && vars.DOCKERHUB_REPOSITORY != ''
-      #   uses: peter-evans/dockerhub-description@v3
-      #   with:
-      #     username: ${{ secrets.DOCKERHUB_USERNAME }}
-      #     password: ${{ secrets.DOCKERHUB_TOKEN }}
-      #     repository: ${{ vars.DOCKERHUB_REPOSITORY }}
-      #     short-description: ${{ github.event.repository.description }}

docs/help/FAQ.md

@@ -196,6 +196,22 @@ tips to reduce the problem:
     === "12GB VRAM GPU"
 
         This should be sufficient to generate larger images up to about 1280x1280.
+		
+## Checkpoint Models Load Slowly or Use Too Much RAM
+
+The difference between diffusers models (a folder containing multiple
+subfolders) and checkpoint models (a file ending with .safetensors or
+.ckpt) is that InvokeAI is able to load diffusers models into memory
+incrementally, while checkpoint models must be loaded all at
+once. With very large models, or systems with limited RAM, you may
+experience slowdowns and other memory-related issues when loading
+checkpoint models.
+
+To solve this, go to the Model Manager tab (the cube), select the
+checkpoint model that's giving you trouble, and press the "Convert"
+button in the upper right of your browser window. This will conver the
+checkpoint into a diffusers model, after which loading should be
+faster and less memory-intensive.
 
 ## Memory Leak (Linux)
 

invokeai/app/api/routers/model_manager.py

@@ -3,8 +3,10 @@
 
 import io
 import pathlib
+import shutil
 import traceback
 from copy import deepcopy
+from enum import Enum
 from tempfile import TemporaryDirectory
 from typing import List, Optional, Type
 
@@ -17,6 +19,7 @@ from starlette.exceptions import HTTPException
 from typing_extensions import Annotated
 
 from invokeai.app.api.dependencies import ApiDependencies
+from invokeai.app.services.config import get_config
 from invokeai.app.services.model_images.model_images_common import ModelImageFileNotFoundException
 from invokeai.app.services.model_install.model_install_common import ModelInstallJob
 from invokeai.app.services.model_records import (
@@ -31,6 +34,7 @@ from invokeai.backend.model_manager.config import (
     ModelFormat,
     ModelType,
 )
+from invokeai.backend.model_manager.load.model_cache.model_cache_base import CacheStats
 from invokeai.backend.model_manager.metadata.fetch.huggingface import HuggingFaceMetadataFetch
 from invokeai.backend.model_manager.metadata.metadata_base import ModelMetadataWithFiles, UnknownMetadataException
 from invokeai.backend.model_manager.search import ModelSearch
@@ -50,6 +54,13 @@ class ModelsList(BaseModel):
     model_config = ConfigDict(use_enum_values=True)
 
 
+class CacheType(str, Enum):
+    """Cache type - one of vram or ram."""
+
+    RAM = "RAM"
+    VRAM = "VRAM"
+
+
 def add_cover_image_to_model_config(config: AnyModelConfig, dependencies: Type[ApiDependencies]) -> AnyModelConfig:
     """Add a cover image URL to a model configuration."""
     cover_image = dependencies.invoker.services.model_images.get_url(config.key)
@@ -797,3 +808,83 @@ async def get_starter_models() -> list[StarterModel]:
         model.dependencies = missing_deps
 
     return starter_models
+
+
+@model_manager_router.get(
+    "/model_cache",
+    operation_id="get_cache_size",
+    response_model=float,
+    summary="Get maximum size of model manager RAM or VRAM cache.",
+)
+async def get_cache_size(cache_type: CacheType = Query(description="The cache type", default=CacheType.RAM)) -> float:
+    """Return the current RAM or VRAM cache size setting (in GB)."""
+    cache = ApiDependencies.invoker.services.model_manager.load.ram_cache
+    value = 0.0
+    if cache_type == CacheType.RAM:
+        value = cache.max_cache_size
+    elif cache_type == CacheType.VRAM:
+        value = cache.max_vram_cache_size
+    return value
+
+
+@model_manager_router.put(
+    "/model_cache",
+    operation_id="set_cache_size",
+    response_model=float,
+    summary="Set maximum size of model manager RAM or VRAM cache, optionally writing new value out to invokeai.yaml config file.",
+)
+async def set_cache_size(
+    value: float = Query(description="The new value for the maximum cache size"),
+    cache_type: CacheType = Query(description="The cache type", default=CacheType.RAM),
+    persist: bool = Query(description="Write new value out to invokeai.yaml", default=False),
+) -> float:
+    """Set the current RAM or VRAM cache size setting (in GB). ."""
+    cache = ApiDependencies.invoker.services.model_manager.load.ram_cache
+    app_config = get_config()
+    # Record initial state.
+    vram_old = app_config.vram
+    ram_old = app_config.ram
+
+    # Prepare target state.
+    vram_new = vram_old
+    ram_new = ram_old
+    if cache_type == CacheType.RAM:
+        ram_new = value
+    elif cache_type == CacheType.VRAM:
+        vram_new = value
+    else:
+        raise ValueError(f"Unexpected {cache_type=}.")
+
+    config_path = app_config.config_file_path
+    new_config_path = config_path.with_suffix(".yaml.new")
+
+    try:
+        # Try to apply the target state.
+        cache.max_vram_cache_size = vram_new
+        cache.max_cache_size = ram_new
+        app_config.ram = ram_new
+        app_config.vram = vram_new
+        if persist:
+            app_config.write_file(new_config_path)
+            shutil.move(new_config_path, config_path)
+    except Exception as e:
+        # If there was a failure, restore the initial state.
+        cache.max_cache_size = ram_old
+        cache.max_vram_cache_size = vram_old
+        app_config.ram = ram_old
+        app_config.vram = vram_old
+
+        raise RuntimeError("Failed to update cache size") from e
+    return value
+
+
+@model_manager_router.get(
+    "/stats",
+    operation_id="get_stats",
+    response_model=Optional[CacheStats],
+    summary="Get model manager RAM cache performance statistics.",
+)
+async def get_stats() -> Optional[CacheStats]:
+    """Return performance statistics on the model manager's RAM cache. Will return null if no models have been loaded."""
+
+    return ApiDependencies.invoker.services.model_manager.load.ram_cache.stats

invokeai/app/invocations/denoise_latents.py

@@ -185,7 +185,7 @@ class DenoiseLatentsInvocation(BaseInvocation):
     )
     denoise_mask: Optional[DenoiseMaskField] = InputField(
         default=None,
-        description=FieldDescriptions.mask,
+        description=FieldDescriptions.denoise_mask,
         input=Input.Connection,
         ui_order=8,
     )

invokeai/app/invocations/fields.py

@@ -45,11 +45,13 @@ class UIType(str, Enum, metaclass=MetaEnum):
     SDXLRefinerModel = "SDXLRefinerModelField"
     ONNXModel = "ONNXModelField"
     VAEModel = "VAEModelField"
+    FluxVAEModel = "FluxVAEModelField"
     LoRAModel = "LoRAModelField"
     ControlNetModel = "ControlNetModelField"
     IPAdapterModel = "IPAdapterModelField"
     T2IAdapterModel = "T2IAdapterModelField"
     T5EncoderModel = "T5EncoderModelField"
+    CLIPEmbedModel = "CLIPEmbedModelField"
     SpandrelImageToImageModel = "SpandrelImageToImageModelField"
     # endregion
 
@@ -128,6 +130,7 @@ class FieldDescriptions:
     noise = "Noise tensor"
     clip = "CLIP (tokenizer, text encoder, LoRAs) and skipped layer count"
     t5_encoder = "T5 tokenizer and text encoder"
+    clip_embed_model = "CLIP Embed loader"
     unet = "UNet (scheduler, LoRAs)"
     transformer = "Transformer"
     vae = "VAE"
@@ -178,7 +181,7 @@ class FieldDescriptions:
     )
     num_1 = "The first number"
     num_2 = "The second number"
-    mask = "The mask to use for the operation"
+    denoise_mask = "A mask of the region to apply the denoising process to."
     board = "The board to save the image to"
     image = "The image to process"
     tile_size = "Tile size"

invokeai/app/invocations/flux_denoise.py

@@ -0,0 +1,249 @@
+from typing import Callable, Optional
+
+import torch
+import torchvision.transforms as tv_transforms
+from torchvision.transforms.functional import resize as tv_resize
+
+from invokeai.app.invocations.baseinvocation import BaseInvocation, Classification, invocation
+from invokeai.app.invocations.fields import (
+    DenoiseMaskField,
+    FieldDescriptions,
+    FluxConditioningField,
+    Input,
+    InputField,
+    LatentsField,
+    WithBoard,
+    WithMetadata,
+)
+from invokeai.app.invocations.model import TransformerField
+from invokeai.app.invocations.primitives import LatentsOutput
+from invokeai.app.services.shared.invocation_context import InvocationContext
+from invokeai.backend.flux.denoise import denoise
+from invokeai.backend.flux.inpaint_extension import InpaintExtension
+from invokeai.backend.flux.model import Flux
+from invokeai.backend.flux.sampling_utils import (
+    clip_timestep_schedule,
+    generate_img_ids,
+    get_noise,
+    get_schedule,
+    pack,
+    unpack,
+)
+from invokeai.backend.stable_diffusion.diffusers_pipeline import PipelineIntermediateState
+from invokeai.backend.stable_diffusion.diffusion.conditioning_data import FLUXConditioningInfo
+from invokeai.backend.util.devices import TorchDevice
+
+
+@invocation(
+    "flux_denoise",
+    title="FLUX Denoise",
+    tags=["image", "flux"],
+    category="image",
+    version="1.0.0",
+    classification=Classification.Prototype,
+)
+class FluxDenoiseInvocation(BaseInvocation, WithMetadata, WithBoard):
+    """Run denoising process with a FLUX transformer model."""
+
+    # If latents is provided, this means we are doing image-to-image.
+    latents: Optional[LatentsField] = InputField(
+        default=None,
+        description=FieldDescriptions.latents,
+        input=Input.Connection,
+    )
+    # denoise_mask is used for image-to-image inpainting. Only the masked region is modified.
+    denoise_mask: Optional[DenoiseMaskField] = InputField(
+        default=None,
+        description=FieldDescriptions.denoise_mask,
+        input=Input.Connection,
+    )
+    denoising_start: float = InputField(
+        default=0.0,
+        ge=0,
+        le=1,
+        description=FieldDescriptions.denoising_start,
+    )
+    denoising_end: float = InputField(default=1.0, ge=0, le=1, description=FieldDescriptions.denoising_end)
+    transformer: TransformerField = InputField(
+        description=FieldDescriptions.flux_model,
+        input=Input.Connection,
+        title="Transformer",
+    )
+    positive_text_conditioning: FluxConditioningField = InputField(
+        description=FieldDescriptions.positive_cond, input=Input.Connection
+    )
+    width: int = InputField(default=1024, multiple_of=16, description="Width of the generated image.")
+    height: int = InputField(default=1024, multiple_of=16, description="Height of the generated image.")
+    num_steps: int = InputField(
+        default=4, description="Number of diffusion steps. Recommended values are schnell: 4, dev: 50."
+    )
+    guidance: float = InputField(
+        default=4.0,
+        description="The guidance strength. Higher values adhere more strictly to the prompt, and will produce less diverse images. FLUX dev only, ignored for schnell.",
+    )
+    seed: int = InputField(default=0, description="Randomness seed for reproducibility.")
+
+    @torch.no_grad()
+    def invoke(self, context: InvocationContext) -> LatentsOutput:
+        latents = self._run_diffusion(context)
+        latents = latents.detach().to("cpu")
+
+        name = context.tensors.save(tensor=latents)
+        return LatentsOutput.build(latents_name=name, latents=latents, seed=None)
+
+    def _run_diffusion(
+        self,
+        context: InvocationContext,
+    ):
+        inference_dtype = torch.bfloat16
+
+        # Load the conditioning data.
+        cond_data = context.conditioning.load(self.positive_text_conditioning.conditioning_name)
+        assert len(cond_data.conditionings) == 1
+        flux_conditioning = cond_data.conditionings[0]
+        assert isinstance(flux_conditioning, FLUXConditioningInfo)
+        flux_conditioning = flux_conditioning.to(dtype=inference_dtype)
+        t5_embeddings = flux_conditioning.t5_embeds
+        clip_embeddings = flux_conditioning.clip_embeds
+
+        # Load the input latents, if provided.
+        init_latents = context.tensors.load(self.latents.latents_name) if self.latents else None
+        if init_latents is not None:
+            init_latents = init_latents.to(device=TorchDevice.choose_torch_device(), dtype=inference_dtype)
+
+        # Prepare input noise.
+        noise = get_noise(
+            num_samples=1,
+            height=self.height,
+            width=self.width,
+            device=TorchDevice.choose_torch_device(),
+            dtype=inference_dtype,
+            seed=self.seed,
+        )
+
+        transformer_info = context.models.load(self.transformer.transformer)
+        is_schnell = "schnell" in transformer_info.config.config_path
+
+        # Calculate the timestep schedule.
+        image_seq_len = noise.shape[-1] * noise.shape[-2] // 4
+        timesteps = get_schedule(
+            num_steps=self.num_steps,
+            image_seq_len=image_seq_len,
+            shift=not is_schnell,
+        )
+
+        # Clip the timesteps schedule based on denoising_start and denoising_end.
+        timesteps = clip_timestep_schedule(timesteps, self.denoising_start, self.denoising_end)
+
+        # Prepare input latent image.
+        if init_latents is not None:
+            # If init_latents is provided, we are doing image-to-image.
+
+            if is_schnell:
+                context.logger.warning(
+                    "Running image-to-image with a FLUX schnell model. This is not recommended. The results are likely "
+                    "to be poor. Consider using a FLUX dev model instead."
+                )
+
+            # Noise the orig_latents by the appropriate amount for the first timestep.
+            t_0 = timesteps[0]
+            x = t_0 * noise + (1.0 - t_0) * init_latents
+        else:
+            # init_latents are not provided, so we are not doing image-to-image (i.e. we are starting from pure noise).
+            if self.denoising_start > 1e-5:
+                raise ValueError("denoising_start should be 0 when initial latents are not provided.")
+
+            x = noise
+
+        # If len(timesteps) == 1, then short-circuit. We are just noising the input latents, but not taking any
+        # denoising steps.
+        if len(timesteps) <= 1:
+            return x
+
+        inpaint_mask = self._prep_inpaint_mask(context, x)
+
+        b, _c, h, w = x.shape
+        img_ids = generate_img_ids(h=h, w=w, batch_size=b, device=x.device, dtype=x.dtype)
+
+        bs, t5_seq_len, _ = t5_embeddings.shape
+        txt_ids = torch.zeros(bs, t5_seq_len, 3, dtype=inference_dtype, device=TorchDevice.choose_torch_device())
+
+        # Pack all latent tensors.
+        init_latents = pack(init_latents) if init_latents is not None else None
+        inpaint_mask = pack(inpaint_mask) if inpaint_mask is not None else None
+        noise = pack(noise)
+        x = pack(x)
+
+        # Now that we have 'packed' the latent tensors, verify that we calculated the image_seq_len correctly.
+        assert image_seq_len == x.shape[1]
+
+        # Prepare inpaint extension.
+        inpaint_extension: InpaintExtension | None = None
+        if inpaint_mask is not None:
+            assert init_latents is not None
+            inpaint_extension = InpaintExtension(
+                init_latents=init_latents,
+                inpaint_mask=inpaint_mask,
+                noise=noise,
+            )
+
+        with transformer_info as transformer:
+            assert isinstance(transformer, Flux)
+
+            x = denoise(
+                model=transformer,
+                img=x,
+                img_ids=img_ids,
+                txt=t5_embeddings,
+                txt_ids=txt_ids,
+                vec=clip_embeddings,
+                timesteps=timesteps,
+                step_callback=self._build_step_callback(context),
+                guidance=self.guidance,
+                inpaint_extension=inpaint_extension,
+            )
+
+        x = unpack(x.float(), self.height, self.width)
+        return x
+
+    def _prep_inpaint_mask(self, context: InvocationContext, latents: torch.Tensor) -> torch.Tensor | None:
+        """Prepare the inpaint mask.
+
+        - Loads the mask
+        - Resizes if necessary
+        - Casts to same device/dtype as latents
+        - Expands mask to the same shape as latents so that they line up after 'packing'
+
+        Args:
+            context (InvocationContext): The invocation context, for loading the inpaint mask.
+            latents (torch.Tensor): A latent image tensor. In 'unpacked' format. Used to determine the target shape,
+                device, and dtype for the inpaint mask.
+
+        Returns:
+            torch.Tensor | None: Inpaint mask.
+        """
+        if self.denoise_mask is None:
+            return None
+
+        mask = context.tensors.load(self.denoise_mask.mask_name)
+
+        _, _, latent_height, latent_width = latents.shape
+        mask = tv_resize(
+            img=mask,
+            size=[latent_height, latent_width],
+            interpolation=tv_transforms.InterpolationMode.BILINEAR,
+            antialias=False,
+        )
+
+        mask = mask.to(device=latents.device, dtype=latents.dtype)
+
+        # Expand the inpaint mask to the same shape as `latents` so that when we 'pack' `mask` it lines up with
+        # `latents`.
+        return mask.expand_as(latents)
+
+    def _build_step_callback(self, context: InvocationContext) -> Callable[[PipelineIntermediateState], None]:
+        def step_callback(state: PipelineIntermediateState) -> None:
+            state.latents = unpack(state.latents.float(), self.height, self.width).squeeze()
+            context.util.flux_step_callback(state)
+
+        return step_callback

invokeai/app/invocations/flux_text_encoder.py

@@ -40,7 +40,10 @@ class FluxTextEncoderInvocation(BaseInvocation):
 
     @torch.no_grad()
     def invoke(self, context: InvocationContext) -> FluxConditioningOutput:
-        t5_embeddings, clip_embeddings = self._encode_prompt(context)
+        # Note: The T5 and CLIP encoding are done in separate functions to ensure that all model references are locally
+        # scoped. This ensures that the T5 model can be freed and gc'd before loading the CLIP model (if necessary).
+        t5_embeddings = self._t5_encode(context)
+        clip_embeddings = self._clip_encode(context)
         conditioning_data = ConditioningFieldData(
             conditionings=[FLUXConditioningInfo(clip_embeds=clip_embeddings, t5_embeds=t5_embeddings)]
         )
@@ -48,12 +51,7 @@ class FluxTextEncoderInvocation(BaseInvocation):
         conditioning_name = context.conditioning.save(conditioning_data)
         return FluxConditioningOutput.build(conditioning_name)
 
-    def _encode_prompt(self, context: InvocationContext) -> tuple[torch.Tensor, torch.Tensor]:
-        # Load CLIP.
-        clip_tokenizer_info = context.models.load(self.clip.tokenizer)
-        clip_text_encoder_info = context.models.load(self.clip.text_encoder)
-
-        # Load T5.
+    def _t5_encode(self, context: InvocationContext) -> torch.Tensor:
         t5_tokenizer_info = context.models.load(self.t5_encoder.tokenizer)
         t5_text_encoder_info = context.models.load(self.t5_encoder.text_encoder)
 
@@ -70,6 +68,15 @@ class FluxTextEncoderInvocation(BaseInvocation):
 
             prompt_embeds = t5_encoder(prompt)
 
+        assert isinstance(prompt_embeds, torch.Tensor)
+        return prompt_embeds
+
+    def _clip_encode(self, context: InvocationContext) -> torch.Tensor:
+        clip_tokenizer_info = context.models.load(self.clip.tokenizer)
+        clip_text_encoder_info = context.models.load(self.clip.text_encoder)
+
+        prompt = [self.prompt]
+
         with (
             clip_text_encoder_info as clip_text_encoder,
             clip_tokenizer_info as clip_tokenizer,
@@ -81,6 +88,5 @@ class FluxTextEncoderInvocation(BaseInvocation):
 
             pooled_prompt_embeds = clip_encoder(prompt)
 
-        assert isinstance(prompt_embeds, torch.Tensor)
         assert isinstance(pooled_prompt_embeds, torch.Tensor)
-        return prompt_embeds, pooled_prompt_embeds
+        return pooled_prompt_embeds

invokeai/app/invocations/flux_text_to_image.py

@@ -1,172 +0,0 @@
-import torch
-from einops import rearrange
-from PIL import Image
-
-from invokeai.app.invocations.baseinvocation import BaseInvocation, Classification, invocation
-from invokeai.app.invocations.fields import (
-    FieldDescriptions,
-    FluxConditioningField,
-    Input,
-    InputField,
-    WithBoard,
-    WithMetadata,
-)
-from invokeai.app.invocations.model import TransformerField, VAEField
-from invokeai.app.invocations.primitives import ImageOutput
-from invokeai.app.services.session_processor.session_processor_common import CanceledException
-from invokeai.app.services.shared.invocation_context import InvocationContext
-from invokeai.backend.flux.model import Flux
-from invokeai.backend.flux.modules.autoencoder import AutoEncoder
-from invokeai.backend.flux.sampling import denoise, get_noise, get_schedule, prepare_latent_img_patches, unpack
-from invokeai.backend.stable_diffusion.diffusion.conditioning_data import FLUXConditioningInfo
-from invokeai.backend.util.devices import TorchDevice
-
-
-@invocation(
-    "flux_text_to_image",
-    title="FLUX Text to Image",
-    tags=["image", "flux"],
-    category="image",
-    version="1.0.0",
-    classification=Classification.Prototype,
-)
-class FluxTextToImageInvocation(BaseInvocation, WithMetadata, WithBoard):
-    """Text-to-image generation using a FLUX model."""
-
-    transformer: TransformerField = InputField(
-        description=FieldDescriptions.flux_model,
-        input=Input.Connection,
-        title="Transformer",
-    )
-    vae: VAEField = InputField(
-        description=FieldDescriptions.vae,
-        input=Input.Connection,
-    )
-    positive_text_conditioning: FluxConditioningField = InputField(
-        description=FieldDescriptions.positive_cond, input=Input.Connection
-    )
-    width: int = InputField(default=1024, multiple_of=16, description="Width of the generated image.")
-    height: int = InputField(default=1024, multiple_of=16, description="Height of the generated image.")
-    num_steps: int = InputField(
-        default=4, description="Number of diffusion steps. Recommend values are schnell: 4, dev: 50."
-    )
-    guidance: float = InputField(
-        default=4.0,
-        description="The guidance strength. Higher values adhere more strictly to the prompt, and will produce less diverse images. FLUX dev only, ignored for schnell.",
-    )
-    seed: int = InputField(default=0, description="Randomness seed for reproducibility.")
-
-    @torch.no_grad()
-    def invoke(self, context: InvocationContext) -> ImageOutput:
-        # Load the conditioning data.
-        cond_data = context.conditioning.load(self.positive_text_conditioning.conditioning_name)
-        assert len(cond_data.conditionings) == 1
-        flux_conditioning = cond_data.conditionings[0]
-        assert isinstance(flux_conditioning, FLUXConditioningInfo)
-
-        latents = self._run_diffusion(context, flux_conditioning.clip_embeds, flux_conditioning.t5_embeds)
-        image = self._run_vae_decoding(context, latents)
-        image_dto = context.images.save(image=image)
-        return ImageOutput.build(image_dto)
-
-    def _run_diffusion(
-        self,
-        context: InvocationContext,
-        clip_embeddings: torch.Tensor,
-        t5_embeddings: torch.Tensor,
-    ):
-        transformer_info = context.models.load(self.transformer.transformer)
-        inference_dtype = torch.bfloat16
-
-        # Prepare input noise.
-        x = get_noise(
-            num_samples=1,
-            height=self.height,
-            width=self.width,
-            device=TorchDevice.choose_torch_device(),
-            dtype=inference_dtype,
-            seed=self.seed,
-        )
-
-        img, img_ids = prepare_latent_img_patches(x)
-
-        is_schnell = "schnell" in transformer_info.config.config_path
-
-        timesteps = get_schedule(
-            num_steps=self.num_steps,
-            image_seq_len=img.shape[1],
-            shift=not is_schnell,
-        )
-
-        bs, t5_seq_len, _ = t5_embeddings.shape
-        txt_ids = torch.zeros(bs, t5_seq_len, 3, dtype=inference_dtype, device=TorchDevice.choose_torch_device())
-
-        # HACK(ryand): Manually empty the cache. Currently we don't check the size of the model before loading it from
-        # disk. Since the transformer model is large (24GB), there's a good chance that it will OOM on 32GB RAM systems
-        # if the cache is not empty.
-        context.models._services.model_manager.load.ram_cache.make_room(24 * 2**30)
-
-        with transformer_info as transformer:
-            assert isinstance(transformer, Flux)
-
-            def step_callback() -> None:
-                if context.util.is_canceled():
-                    raise CanceledException
-
-                # TODO: Make this look like the image before re-enabling
-                # latent_image = unpack(img.float(), self.height, self.width)
-                # latent_image = latent_image.squeeze()  # Remove unnecessary dimensions
-                # flattened_tensor = latent_image.reshape(-1)  # Flatten to shape [48*128*128]
-
-                # # Create a new tensor of the required shape [255, 255, 3]
-                # latent_image = flattened_tensor[: 255 * 255 * 3].reshape(255, 255, 3)  # Reshape to RGB format
-
-                # # Convert to a NumPy array and then to a PIL Image
-                # image = Image.fromarray(latent_image.cpu().numpy().astype(np.uint8))
-
-                # (width, height) = image.size
-                # width *= 8
-                # height *= 8
-
-                # dataURL = image_to_dataURL(image, image_format="JPEG")
-
-                # # TODO: move this whole function to invocation context to properly reference these variables
-                # context._services.events.emit_invocation_denoise_progress(
-                #     context._data.queue_item,
-                #     context._data.invocation,
-                #     state,
-                #     ProgressImage(dataURL=dataURL, width=width, height=height),
-                # )
-
-            x = denoise(
-                model=transformer,
-                img=img,
-                img_ids=img_ids,
-                txt=t5_embeddings,
-                txt_ids=txt_ids,
-                vec=clip_embeddings,
-                timesteps=timesteps,
-                step_callback=step_callback,
-                guidance=self.guidance,
-            )
-
-        x = unpack(x.float(), self.height, self.width)
-
-        return x
-
-    def _run_vae_decoding(
-        self,
-        context: InvocationContext,
-        latents: torch.Tensor,
-    ) -> Image.Image:
-        vae_info = context.models.load(self.vae.vae)
-        with vae_info as vae:
-            assert isinstance(vae, AutoEncoder)
-            latents = latents.to(dtype=TorchDevice.choose_torch_dtype())
-            img = vae.decode(latents)
-
-        img = img.clamp(-1, 1)
-        img = rearrange(img[0], "c h w -> h w c")
-        img_pil = Image.fromarray((127.5 * (img + 1.0)).byte().cpu().numpy())
-
-        return img_pil

invokeai/app/invocations/flux_vae_decode.py

@@ -0,0 +1,60 @@
+import torch
+from einops import rearrange
+from PIL import Image
+
+from invokeai.app.invocations.baseinvocation import BaseInvocation, invocation
+from invokeai.app.invocations.fields import (
+    FieldDescriptions,
+    Input,
+    InputField,
+    LatentsField,
+    WithBoard,
+    WithMetadata,
+)
+from invokeai.app.invocations.model import VAEField
+from invokeai.app.invocations.primitives import ImageOutput
+from invokeai.app.services.shared.invocation_context import InvocationContext
+from invokeai.backend.flux.modules.autoencoder import AutoEncoder
+from invokeai.backend.model_manager.load.load_base import LoadedModel
+from invokeai.backend.util.devices import TorchDevice
+
+
+@invocation(
+    "flux_vae_decode",
+    title="FLUX Latents to Image",
+    tags=["latents", "image", "vae", "l2i", "flux"],
+    category="latents",
+    version="1.0.0",
+)
+class FluxVaeDecodeInvocation(BaseInvocation, WithMetadata, WithBoard):
+    """Generates an image from latents."""
+
+    latents: LatentsField = InputField(
+        description=FieldDescriptions.latents,
+        input=Input.Connection,
+    )
+    vae: VAEField = InputField(
+        description=FieldDescriptions.vae,
+        input=Input.Connection,
+    )
+
+    def _vae_decode(self, vae_info: LoadedModel, latents: torch.Tensor) -> Image.Image:
+        with vae_info as vae:
+            assert isinstance(vae, AutoEncoder)
+            latents = latents.to(device=TorchDevice.choose_torch_device(), dtype=TorchDevice.choose_torch_dtype())
+            img = vae.decode(latents)
+
+        img = img.clamp(-1, 1)
+        img = rearrange(img[0], "c h w -> h w c")  # noqa: F821
+        img_pil = Image.fromarray((127.5 * (img + 1.0)).byte().cpu().numpy())
+        return img_pil
+
+    @torch.no_grad()
+    def invoke(self, context: InvocationContext) -> ImageOutput:
+        latents = context.tensors.load(self.latents.latents_name)
+        vae_info = context.models.load(self.vae.vae)
+        image = self._vae_decode(vae_info=vae_info, latents=latents)
+
+        TorchDevice.empty_cache()
+        image_dto = context.images.save(image=image)
+        return ImageOutput.build(image_dto)

invokeai/app/invocations/flux_vae_encode.py

@@ -0,0 +1,67 @@
+import einops
+import torch
+
+from invokeai.app.invocations.baseinvocation import BaseInvocation, invocation
+from invokeai.app.invocations.fields import (
+    FieldDescriptions,
+    ImageField,
+    Input,
+    InputField,
+)
+from invokeai.app.invocations.model import VAEField
+from invokeai.app.invocations.primitives import LatentsOutput
+from invokeai.app.services.shared.invocation_context import InvocationContext
+from invokeai.backend.flux.modules.autoencoder import AutoEncoder
+from invokeai.backend.model_manager import LoadedModel
+from invokeai.backend.stable_diffusion.diffusers_pipeline import image_resized_to_grid_as_tensor
+from invokeai.backend.util.devices import TorchDevice
+
+
+@invocation(
+    "flux_vae_encode",
+    title="FLUX Image to Latents",
+    tags=["latents", "image", "vae", "i2l", "flux"],
+    category="latents",
+    version="1.0.0",
+)
+class FluxVaeEncodeInvocation(BaseInvocation):
+    """Encodes an image into latents."""
+
+    image: ImageField = InputField(
+        description="The image to encode.",
+    )
+    vae: VAEField = InputField(
+        description=FieldDescriptions.vae,
+        input=Input.Connection,
+    )
+
+    @staticmethod
+    def vae_encode(vae_info: LoadedModel, image_tensor: torch.Tensor) -> torch.Tensor:
+        # TODO(ryand): Expose seed parameter at the invocation level.
+        # TODO(ryand): Write a util function for generating random tensors that is consistent across devices / dtypes.
+        # There's a starting point in get_noise(...), but it needs to be extracted and generalized. This function
+        # should be used for VAE encode sampling.
+        generator = torch.Generator(device=TorchDevice.choose_torch_device()).manual_seed(0)
+        with vae_info as vae:
+            assert isinstance(vae, AutoEncoder)
+            image_tensor = image_tensor.to(
+                device=TorchDevice.choose_torch_device(), dtype=TorchDevice.choose_torch_dtype()
+            )
+            latents = vae.encode(image_tensor, sample=True, generator=generator)
+            return latents
+
+    @torch.no_grad()
+    def invoke(self, context: InvocationContext) -> LatentsOutput:
+        image = context.images.get_pil(self.image.image_name)
+
+        vae_info = context.models.load(self.vae.vae)
+
+        image_tensor = image_resized_to_grid_as_tensor(image.convert("RGB"))
+        if image_tensor.dim() == 3:
+            image_tensor = einops.rearrange(image_tensor, "c h w -> 1 c h w")
+
+        latents = self.vae_encode(vae_info=vae_info, image_tensor=image_tensor)
+
+        latents = latents.to("cpu")
+        name = context.tensors.save(tensor=latents)
+        return LatentsOutput.build(latents_name=name, latents=latents, seed=None)

invokeai/app/invocations/mask.py

@@ -126,7 +126,7 @@ class ImageMaskToTensorInvocation(BaseInvocation, WithMetadata):
     title="Tensor Mask to Image",
     tags=["mask"],
     category="mask",
-    version="1.0.0",
+    version="1.1.0",
 )
 class MaskTensorToImageInvocation(BaseInvocation, WithMetadata, WithBoard):
     """Convert a mask tensor to an image."""
@@ -135,6 +135,11 @@ class MaskTensorToImageInvocation(BaseInvocation, WithMetadata, WithBoard):
 
     def invoke(self, context: InvocationContext) -> ImageOutput:
         mask = context.tensors.load(self.mask.tensor_name)
+
+        # Squeeze the channel dimension if it exists.
+        if mask.dim() == 3:
+            mask = mask.squeeze(0)
+
         # Ensure that the mask is binary.
         if mask.dtype != torch.bool:
             mask = mask > 0.5

invokeai/app/invocations/model.py

@@ -157,7 +157,7 @@ class FluxModelLoaderOutput(BaseInvocationOutput):
     title="Flux Main Model",
     tags=["model", "flux"],
     category="model",
-    version="1.0.3",
+    version="1.0.4",
     classification=Classification.Prototype,
 )
 class FluxModelLoaderInvocation(BaseInvocation):
@@ -169,23 +169,35 @@ class FluxModelLoaderInvocation(BaseInvocation):
         input=Input.Direct,
     )
 
-    t5_encoder: ModelIdentifierField = InputField(
-        description=FieldDescriptions.t5_encoder,
-        ui_type=UIType.T5EncoderModel,
+    t5_encoder_model: ModelIdentifierField = InputField(
+        description=FieldDescriptions.t5_encoder, ui_type=UIType.T5EncoderModel, input=Input.Direct, title="T5 Encoder"
+    )
+
+    clip_embed_model: ModelIdentifierField = InputField(
+        description=FieldDescriptions.clip_embed_model,
+        ui_type=UIType.CLIPEmbedModel,
         input=Input.Direct,
+        title="CLIP Embed",
+    )
+
+    vae_model: ModelIdentifierField = InputField(
+        description=FieldDescriptions.vae_model, ui_type=UIType.FluxVAEModel, title="VAE"
     )
 
     def invoke(self, context: InvocationContext) -> FluxModelLoaderOutput:
-        model_key = self.model.key
+        for key in [self.model.key, self.t5_encoder_model.key, self.clip_embed_model.key, self.vae_model.key]:
+            if not context.models.exists(key):
+                raise ValueError(f"Unknown model: {key}")
+
+        transformer = self.model.model_copy(update={"submodel_type": SubModelType.Transformer})
+        vae = self.vae_model.model_copy(update={"submodel_type": SubModelType.VAE})
+
+        tokenizer = self.clip_embed_model.model_copy(update={"submodel_type": SubModelType.Tokenizer})
+        clip_encoder = self.clip_embed_model.model_copy(update={"submodel_type": SubModelType.TextEncoder})
+
+        tokenizer2 = self.t5_encoder_model.model_copy(update={"submodel_type": SubModelType.Tokenizer2})
+        t5_encoder = self.t5_encoder_model.model_copy(update={"submodel_type": SubModelType.TextEncoder2})
 
-        if not context.models.exists(model_key):
-            raise ValueError(f"Unknown model: {model_key}")
-        transformer = self._get_model(context, SubModelType.Transformer)
-        tokenizer = self._get_model(context, SubModelType.Tokenizer)
-        tokenizer2 = self._get_model(context, SubModelType.Tokenizer2)
-        clip_encoder = self._get_model(context, SubModelType.TextEncoder)
-        t5_encoder = self._get_model(context, SubModelType.TextEncoder2)
-        vae = self._get_model(context, SubModelType.VAE)
         transformer_config = context.models.get_config(transformer)
         assert isinstance(transformer_config, CheckpointConfigBase)
 
@@ -197,52 +209,6 @@ class FluxModelLoaderInvocation(BaseInvocation):
             max_seq_len=max_seq_lengths[transformer_config.config_path],
         )
 
-    def _get_model(self, context: InvocationContext, submodel: SubModelType) -> ModelIdentifierField:
-        match submodel:
-            case SubModelType.Transformer:
-                return self.model.model_copy(update={"submodel_type": SubModelType.Transformer})
-            case SubModelType.VAE:
-                return self._pull_model_from_mm(
-                    context,
-                    SubModelType.VAE,
-                    "FLUX.1-schnell_ae",
-                    ModelType.VAE,
-                    BaseModelType.Flux,
-                )
-            case submodel if submodel in [SubModelType.Tokenizer, SubModelType.TextEncoder]:
-                return self._pull_model_from_mm(
-                    context,
-                    submodel,
-                    "clip-vit-large-patch14",
-                    ModelType.CLIPEmbed,
-                    BaseModelType.Any,
-                )
-            case submodel if submodel in [SubModelType.Tokenizer2, SubModelType.TextEncoder2]:
-                return self._pull_model_from_mm(
-                    context,
-                    submodel,
-                    self.t5_encoder.name,
-                    ModelType.T5Encoder,
-                    BaseModelType.Any,
-                )
-            case _:
-                raise Exception(f"{submodel.value} is not a supported submodule for a flux model")
-
-    def _pull_model_from_mm(
-        self,
-        context: InvocationContext,
-        submodel: SubModelType,
-        name: str,
-        type: ModelType,
-        base: BaseModelType,
-    ):
-        if models := context.models.search_by_attrs(name=name, base=base, type=type):
-            if len(models) != 1:
-                raise Exception(f"Multiple models detected for selected model with name {name}")
-            return ModelIdentifierField.from_config(models[0]).model_copy(update={"submodel_type": submodel})
-        else:
-            raise ValueError(f"Please install the {base}:{type} model named {name} via starter models")
-
 
 @invocation(
     "main_model_loader",

invokeai/app/services/events/events_common.py

@@ -88,7 +88,8 @@ class QueueItemEventBase(QueueEventBase):
 
     item_id: int = Field(description="The ID of the queue item")
     batch_id: str = Field(description="The ID of the queue batch")
-    origin: str | None = Field(default=None, description="The origin of the batch")
+    origin: str | None = Field(default=None, description="The origin of the queue item")
+    destination: str | None = Field(default=None, description="The destination of the queue item")
 
 
 class InvocationEventBase(QueueItemEventBase):
@@ -114,6 +115,7 @@ class InvocationStartedEvent(InvocationEventBase):
             item_id=queue_item.item_id,
             batch_id=queue_item.batch_id,
             origin=queue_item.origin,
+            destination=queue_item.destination,
             session_id=queue_item.session_id,
             invocation=invocation,
             invocation_source_id=queue_item.session.prepared_source_mapping[invocation.id],
@@ -148,6 +150,7 @@ class InvocationDenoiseProgressEvent(InvocationEventBase):
             item_id=queue_item.item_id,
             batch_id=queue_item.batch_id,
             origin=queue_item.origin,
+            destination=queue_item.destination,
             session_id=queue_item.session_id,
             invocation=invocation,
             invocation_source_id=queue_item.session.prepared_source_mapping[invocation.id],
@@ -186,6 +189,7 @@ class InvocationCompleteEvent(InvocationEventBase):
             item_id=queue_item.item_id,
             batch_id=queue_item.batch_id,
             origin=queue_item.origin,
+            destination=queue_item.destination,
             session_id=queue_item.session_id,
             invocation=invocation,
             invocation_source_id=queue_item.session.prepared_source_mapping[invocation.id],
@@ -219,6 +223,7 @@ class InvocationErrorEvent(InvocationEventBase):
             item_id=queue_item.item_id,
             batch_id=queue_item.batch_id,
             origin=queue_item.origin,
+            destination=queue_item.destination,
             session_id=queue_item.session_id,
             invocation=invocation,
             invocation_source_id=queue_item.session.prepared_source_mapping[invocation.id],
@@ -257,6 +262,7 @@ class QueueItemStatusChangedEvent(QueueItemEventBase):
             item_id=queue_item.item_id,
             batch_id=queue_item.batch_id,
             origin=queue_item.origin,
+            destination=queue_item.destination,
             session_id=queue_item.session_id,
             status=queue_item.status,
             error_type=queue_item.error_type,

invokeai/app/services/model_install/model_install_common.py

@@ -103,7 +103,7 @@ class HFModelSource(StringLikeSource):
         if self.variant:
             base += f":{self.variant or ''}"
         if self.subfolder:
-            base += f":{self.subfolder}"
+            base += f"::{self.subfolder.as_posix()}"
         return base
 
 

invokeai/app/services/session_queue/session_queue_common.py

@@ -77,7 +77,14 @@ BatchDataCollection: TypeAlias = list[list[BatchDatum]]
 
 class Batch(BaseModel):
     batch_id: str = Field(default_factory=uuid_string, description="The ID of the batch")
-    origin: str | None = Field(default=None, description="The origin of this batch.")
+    origin: str | None = Field(
+        default=None,
+        description="The origin of this queue item. This data is used by the frontend to determine how to handle results.",
+    )
+    destination: str | None = Field(
+        default=None,
+        description="The origin of this queue item. This data is used by the frontend to determine how to handle results",
+    )
     data: Optional[BatchDataCollection] = Field(default=None, description="The batch data collection.")
     graph: Graph = Field(description="The graph to initialize the session with")
     workflow: Optional[WorkflowWithoutID] = Field(
@@ -196,7 +203,14 @@ class SessionQueueItemWithoutGraph(BaseModel):
     status: QUEUE_ITEM_STATUS = Field(default="pending", description="The status of this queue item")
     priority: int = Field(default=0, description="The priority of this queue item")
     batch_id: str = Field(description="The ID of the batch associated with this queue item")
-    origin: str | None = Field(default=None, description="The origin of this queue item. ")
+    origin: str | None = Field(
+        default=None,
+        description="The origin of this queue item. This data is used by the frontend to determine how to handle results.",
+    )
+    destination: str | None = Field(
+        default=None,
+        description="The origin of this queue item. This data is used by the frontend to determine how to handle results",
+    )
     session_id: str = Field(
         description="The ID of the session associated with this queue item. The session doesn't exist in graph_executions until the queue item is executed."
     )
@@ -297,6 +311,7 @@ class BatchStatus(BaseModel):
     queue_id: str = Field(..., description="The ID of the queue")
     batch_id: str = Field(..., description="The ID of the batch")
     origin: str | None = Field(..., description="The origin of the batch")
+    destination: str | None = Field(..., description="The destination of the batch")
     pending: int = Field(..., description="Number of queue items with status 'pending'")
     in_progress: int = Field(..., description="Number of queue items with status 'in_progress'")
     completed: int = Field(..., description="Number of queue items with status 'complete'")
@@ -443,6 +458,7 @@ class SessionQueueValueToInsert(NamedTuple):
     priority: int  # priority
     workflow: Optional[str]  # workflow json
     origin: str | None
+    destination: str | None
 
 
 ValuesToInsert: TypeAlias = list[SessionQueueValueToInsert]
@@ -464,6 +480,7 @@ def prepare_values_to_insert(queue_id: str, batch: Batch, priority: int, max_new
                 priority,  # priority
                 json.dumps(workflow, default=to_jsonable_python) if workflow else None,  # workflow (json)
                 batch.origin,  # origin
+                batch.destination,  # destination
             )
         )
     return values_to_insert

invokeai/app/services/session_queue/session_queue_sqlite.py

@@ -128,8 +128,8 @@ class SqliteSessionQueue(SessionQueueBase):
 
             self.__cursor.executemany(
                 """--sql
-                INSERT INTO session_queue (queue_id, session, session_id, batch_id, field_values, priority, workflow, origin)
-                VALUES (?, ?, ?, ?, ?, ?, ?, ?)
+                INSERT INTO session_queue (queue_id, session, session_id, batch_id, field_values, priority, workflow, origin, destination)
+                VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?)
                 """,
                 values_to_insert,
             )
@@ -579,7 +579,8 @@ class SqliteSessionQueue(SessionQueueBase):
                     session_id,
                     batch_id,
                     queue_id,
-                    origin
+                    origin,
+                    destination
                 FROM session_queue
                 WHERE queue_id = ?
             """
@@ -659,7 +660,7 @@ class SqliteSessionQueue(SessionQueueBase):
             self.__lock.acquire()
             self.__cursor.execute(
                 """--sql
-                SELECT status, count(*), origin
+                SELECT status, count(*), origin, destination
                 FROM session_queue
                 WHERE
                   queue_id = ?
@@ -672,6 +673,7 @@ class SqliteSessionQueue(SessionQueueBase):
             total = sum(row[1] for row in result)
             counts: dict[str, int] = {row[0]: row[1] for row in result}
             origin = result[0]["origin"] if result else None
+            destination = result[0]["destination"] if result else None
         except Exception:
             self.__conn.rollback()
             raise
@@ -681,6 +683,7 @@ class SqliteSessionQueue(SessionQueueBase):
         return BatchStatus(
             batch_id=batch_id,
             origin=origin,
+            destination=destination,
             queue_id=queue_id,
             pending=counts.get("pending", 0),
             in_progress=counts.get("in_progress", 0),

invokeai/app/services/shared/invocation_context.py

@@ -14,7 +14,7 @@ from invokeai.app.services.image_records.image_records_common import ImageCatego
 from invokeai.app.services.images.images_common import ImageDTO
 from invokeai.app.services.invocation_services import InvocationServices
 from invokeai.app.services.model_records.model_records_base import UnknownModelException
-from invokeai.app.util.step_callback import stable_diffusion_step_callback
+from invokeai.app.util.step_callback import flux_step_callback, stable_diffusion_step_callback
 from invokeai.backend.model_manager.config import (
     AnyModel,
     AnyModelConfig,
@@ -557,6 +557,24 @@ class UtilInterface(InvocationContextInterface):
             is_canceled=self.is_canceled,
         )
 
+    def flux_step_callback(self, intermediate_state: PipelineIntermediateState) -> None:
+        """
+        The step callback emits a progress event with the current step, the total number of
+        steps, a preview image, and some other internal metadata.
+
+        This should be called after each denoising step.
+
+        Args:
+            intermediate_state: The intermediate state of the diffusion pipeline.
+        """
+
+        flux_step_callback(
+            context_data=self._data,
+            intermediate_state=intermediate_state,
+            events=self._services.events,
+            is_canceled=self.is_canceled,
+        )
+
 
 class InvocationContext:
     """Provides access to various services and data for the current invocation.

invokeai/app/services/shared/sqlite_migrator/migrations/migration_15.py

@@ -10,9 +10,11 @@ class Migration15Callback:
     def _add_origin_col(self, cursor: sqlite3.Cursor) -> None:
         """
         - Adds `origin` column to the session queue table.
+        - Adds `destination` column to the session queue table.
         """
 
         cursor.execute("ALTER TABLE session_queue ADD COLUMN origin TEXT;")
+        cursor.execute("ALTER TABLE session_queue ADD COLUMN destination TEXT;")
 
 
 def build_migration_15() -> Migration:
@@ -21,6 +23,7 @@ def build_migration_15() -> Migration:
 
     This migration does the following:
         - Adds `origin` column to the session queue table.
+        - Adds `destination` column to the session queue table.
     """
     migration_15 = Migration(
         from_version=14,

invokeai/app/services/workflow_records/default_workflows/FLUX Image to Image.json

@@ -0,0 +1,407 @@
+{
+  "name": "FLUX Image to Image",
+  "author": "InvokeAI",
+  "description": "A simple image-to-image workflow using a FLUX dev model. ",
+  "version": "1.0.4",
+  "contact": "",
+  "tags": "image2image, flux, image-to-image",
+  "notes": "Prerequisite model downloads: T5 Encoder, CLIP-L Encoder, and FLUX VAE. Quantized and un-quantized versions can be found in the starter models tab within your Model Manager. We recommend using FLUX dev models for image-to-image workflows. The image-to-image performance with FLUX schnell models is poor.",
+  "exposedFields": [
+    {
+      "nodeId": "f8d9d7c8-9ed7-4bd7-9e42-ab0e89bfac90",
+      "fieldName": "model"
+    },
+    {
+      "nodeId": "f8d9d7c8-9ed7-4bd7-9e42-ab0e89bfac90",
+      "fieldName": "t5_encoder_model"
+    },
+    {
+      "nodeId": "f8d9d7c8-9ed7-4bd7-9e42-ab0e89bfac90",
+      "fieldName": "clip_embed_model"
+    },
+    {
+      "nodeId": "f8d9d7c8-9ed7-4bd7-9e42-ab0e89bfac90",
+      "fieldName": "vae_model"
+    },
+    {
+      "nodeId": "ace0258f-67d7-4eee-a218-6fff27065214",
+      "fieldName": "denoising_start"
+    },
+    {
+      "nodeId": "01f674f8-b3d1-4df1-acac-6cb8e0bfb63c",
+      "fieldName": "prompt"
+    },
+    {
+      "nodeId": "ace0258f-67d7-4eee-a218-6fff27065214",
+      "fieldName": "num_steps"
+    }
+  ],
+  "meta": {
+    "version": "3.0.0",
+    "category": "default"
+  },
+  "nodes": [
+    {
+      "id": "2981a67c-480f-4237-9384-26b68dbf912b",
+      "type": "invocation",
+      "data": {
+        "id": "2981a67c-480f-4237-9384-26b68dbf912b",
+        "type": "flux_vae_encode",
+        "version": "1.0.0",
+        "label": "",
+        "notes": "",
+        "isOpen": true,
+        "isIntermediate": true,
+        "useCache": true,
+        "inputs": {
+          "image": {
+            "name": "image",
+            "label": "",
+            "value": {
+              "image_name": "8a5c62aa-9335-45d2-9c71-89af9fc1f8d4.png"
+            }
+          },
+          "vae": {
+            "name": "vae",
+            "label": ""
+          }
+        }
+      },
+      "position": {
+        "x": 732.7680166609682,
+        "y": -24.37398171806909
+      }
+    },
+    {
+      "id": "ace0258f-67d7-4eee-a218-6fff27065214",
+      "type": "invocation",
+      "data": {
+        "id": "ace0258f-67d7-4eee-a218-6fff27065214",
+        "type": "flux_denoise",
+        "version": "1.0.0",
+        "label": "",
+        "notes": "",
+        "isOpen": true,
+        "isIntermediate": true,
+        "useCache": true,
+        "inputs": {
+          "board": {
+            "name": "board",
+            "label": ""
+          },
+          "metadata": {
+            "name": "metadata",
+            "label": ""
+          },
+          "latents": {
+            "name": "latents",
+            "label": ""
+          },
+          "denoise_mask": {
+            "name": "denoise_mask",
+            "label": ""
+          },
+          "denoising_start": {
+            "name": "denoising_start",
+            "label": "",
+            "value": 0.04
+          },
+          "denoising_end": {
+            "name": "denoising_end",
+            "label": "",
+            "value": 1
+          },
+          "transformer": {
+            "name": "transformer",
+            "label": ""
+          },
+          "positive_text_conditioning": {
+            "name": "positive_text_conditioning",
+            "label": ""
+          },
+          "width": {
+            "name": "width",
+            "label": "",
+            "value": 1024
+          },
+          "height": {
+            "name": "height",
+            "label": "",
+            "value": 1024
+          },
+          "num_steps": {
+            "name": "num_steps",
+            "label": "Steps (Recommend 30 for Dev, 4 for Schnell)",
+            "value": 30
+          },
+          "guidance": {
+            "name": "guidance",
+            "label": "",
+            "value": 4
+          },
+          "seed": {
+            "name": "seed",
+            "label": "",
+            "value": 0
+          }
+        }
+      },
+      "position": {
+        "x": 1182.8836633018684,
+        "y": -251.38882958913183
+      }
+    },
+    {
+      "id": "7e5172eb-48c1-44db-a770-8fd83e1435d1",
+      "type": "invocation",
+      "data": {
+        "id": "7e5172eb-48c1-44db-a770-8fd83e1435d1",
+        "type": "flux_vae_decode",
+        "version": "1.0.0",
+        "label": "",
+        "notes": "",
+        "isOpen": true,
+        "isIntermediate": false,
+        "useCache": true,
+        "inputs": {
+          "board": {
+            "name": "board",
+            "label": ""
+          },
+          "metadata": {
+            "name": "metadata",
+            "label": ""
+          },
+          "latents": {
+            "name": "latents",
+            "label": ""
+          },
+          "vae": {
+            "name": "vae",
+            "label": ""
+          }
+        }
+      },
+      "position": {
+        "x": 1575.5797431839133,
+        "y": -209.00150975507415
+      }
+    },
+    {
+      "id": "f8d9d7c8-9ed7-4bd7-9e42-ab0e89bfac90",
+      "type": "invocation",
+      "data": {
+        "id": "f8d9d7c8-9ed7-4bd7-9e42-ab0e89bfac90",
+        "type": "flux_model_loader",
+        "version": "1.0.4",
+        "label": "",
+        "notes": "",
+        "isOpen": true,
+        "isIntermediate": true,
+        "useCache": false,
+        "inputs": {
+          "model": {
+            "name": "model",
+            "label": "Model (dev variant recommended for Image-to-Image)"
+          },
+          "t5_encoder_model": {
+            "name": "t5_encoder_model",
+            "label": ""
+          },
+          "clip_embed_model": {
+            "name": "clip_embed_model",
+            "label": "",
+            "value": {
+              "key": "fa23a584-b623-415d-832a-21b5098ff1a1",
+              "hash": "blake3:17c19f0ef941c3b7609a9c94a659ca5364de0be364a91d4179f0e39ba17c3b70",
+              "name": "clip-vit-large-patch14",
+              "base": "any",
+              "type": "clip_embed"
+            }
+          },
+          "vae_model": {
+            "name": "vae_model",
+            "label": "",
+            "value": {
+              "key": "74fc82ba-c0a8-479d-a890-2126f82da758",
+              "hash": "blake3:ce21cb76364aa6e2421311cf4a4b5eb052a76c4f1cd207b50703d8978198a068",
+              "name": "FLUX.1-schnell_ae",
+              "base": "flux",
+              "type": "vae"
+            }
+          }
+        }
+      },
+      "position": {
+        "x": 328.1809894659957,
+        "y": -90.2241133566946
+      }
+    },
+    {
+      "id": "01f674f8-b3d1-4df1-acac-6cb8e0bfb63c",
+      "type": "invocation",
+      "data": {
+        "id": "01f674f8-b3d1-4df1-acac-6cb8e0bfb63c",
+        "type": "flux_text_encoder",
+        "version": "1.0.0",
+        "label": "",
+        "notes": "",
+        "isOpen": true,
+        "isIntermediate": true,
+        "useCache": true,
+        "inputs": {
+          "clip": {
+            "name": "clip",
+            "label": ""
+          },
+          "t5_encoder": {
+            "name": "t5_encoder",
+            "label": ""
+          },
+          "t5_max_seq_len": {
+            "name": "t5_max_seq_len",
+            "label": "T5 Max Seq Len",
+            "value": 256
+          },
+          "prompt": {
+            "name": "prompt",
+            "label": "",
+            "value": "a cat wearing a birthday hat"
+          }
+        }
+      },
+      "position": {
+        "x": 745.8823365057267,
+        "y": -299.60249175851914
+      }
+    },
+    {
+      "id": "4754c534-a5f3-4ad0-9382-7887985e668c",
+      "type": "invocation",
+      "data": {
+        "id": "4754c534-a5f3-4ad0-9382-7887985e668c",
+        "type": "rand_int",
+        "version": "1.0.1",
+        "label": "",
+        "notes": "",
+        "isOpen": true,
+        "isIntermediate": true,
+        "useCache": false,
+        "inputs": {
+          "low": {
+            "name": "low",
+            "label": "",
+            "value": 0
+          },
+          "high": {
+            "name": "high",
+            "label": "",
+            "value": 2147483647
+          }
+        }
+      },
+      "position": {
+        "x": 725.834098928012,
+        "y": 496.2710031089931
+      }
+    }
+  ],
+  "edges": [
+    {
+      "id": "reactflow__edge-2981a67c-480f-4237-9384-26b68dbf912bheight-ace0258f-67d7-4eee-a218-6fff27065214height",
+      "type": "default",
+      "source": "2981a67c-480f-4237-9384-26b68dbf912b",
+      "target": "ace0258f-67d7-4eee-a218-6fff27065214",
+      "sourceHandle": "height",
+      "targetHandle": "height"
+    },
+    {
+      "id": "reactflow__edge-2981a67c-480f-4237-9384-26b68dbf912bwidth-ace0258f-67d7-4eee-a218-6fff27065214width",
+      "type": "default",
+      "source": "2981a67c-480f-4237-9384-26b68dbf912b",
+      "target": "ace0258f-67d7-4eee-a218-6fff27065214",
+      "sourceHandle": "width",
+      "targetHandle": "width"
+    },
+    {
+      "id": "reactflow__edge-2981a67c-480f-4237-9384-26b68dbf912blatents-ace0258f-67d7-4eee-a218-6fff27065214latents",
+      "type": "default",
+      "source": "2981a67c-480f-4237-9384-26b68dbf912b",
+      "target": "ace0258f-67d7-4eee-a218-6fff27065214",
+      "sourceHandle": "latents",
+      "targetHandle": "latents"
+    },
+    {
+      "id": "reactflow__edge-f8d9d7c8-9ed7-4bd7-9e42-ab0e89bfac90vae-2981a67c-480f-4237-9384-26b68dbf912bvae",
+      "type": "default",
+      "source": "f8d9d7c8-9ed7-4bd7-9e42-ab0e89bfac90",
+      "target": "2981a67c-480f-4237-9384-26b68dbf912b",
+      "sourceHandle": "vae",
+      "targetHandle": "vae"
+    },
+    {
+      "id": "reactflow__edge-ace0258f-67d7-4eee-a218-6fff27065214latents-7e5172eb-48c1-44db-a770-8fd83e1435d1latents",
+      "type": "default",
+      "source": "ace0258f-67d7-4eee-a218-6fff27065214",
+      "target": "7e5172eb-48c1-44db-a770-8fd83e1435d1",
+      "sourceHandle": "latents",
+      "targetHandle": "latents"
+    },
+    {
+      "id": "reactflow__edge-4754c534-a5f3-4ad0-9382-7887985e668cvalue-ace0258f-67d7-4eee-a218-6fff27065214seed",
+      "type": "default",
+      "source": "4754c534-a5f3-4ad0-9382-7887985e668c",
+      "target": "ace0258f-67d7-4eee-a218-6fff27065214",
+      "sourceHandle": "value",
+      "targetHandle": "seed"
+    },
+    {
+      "id": "reactflow__edge-f8d9d7c8-9ed7-4bd7-9e42-ab0e89bfac90transformer-ace0258f-67d7-4eee-a218-6fff27065214transformer",
+      "type": "default",
+      "source": "f8d9d7c8-9ed7-4bd7-9e42-ab0e89bfac90",
+      "target": "ace0258f-67d7-4eee-a218-6fff27065214",
+      "sourceHandle": "transformer",
+      "targetHandle": "transformer"
+    },
+    {
+      "id": "reactflow__edge-01f674f8-b3d1-4df1-acac-6cb8e0bfb63cconditioning-ace0258f-67d7-4eee-a218-6fff27065214positive_text_conditioning",
+      "type": "default",
+      "source": "01f674f8-b3d1-4df1-acac-6cb8e0bfb63c",
+      "target": "ace0258f-67d7-4eee-a218-6fff27065214",
+      "sourceHandle": "conditioning",
+      "targetHandle": "positive_text_conditioning"
+    },
+    {
+      "id": "reactflow__edge-f8d9d7c8-9ed7-4bd7-9e42-ab0e89bfac90vae-7e5172eb-48c1-44db-a770-8fd83e1435d1vae",
+      "type": "default",
+      "source": "f8d9d7c8-9ed7-4bd7-9e42-ab0e89bfac90",
+      "target": "7e5172eb-48c1-44db-a770-8fd83e1435d1",
+      "sourceHandle": "vae",
+      "targetHandle": "vae"
+    },
+    {
+      "id": "reactflow__edge-f8d9d7c8-9ed7-4bd7-9e42-ab0e89bfac90max_seq_len-01f674f8-b3d1-4df1-acac-6cb8e0bfb63ct5_max_seq_len",
+      "type": "default",
+      "source": "f8d9d7c8-9ed7-4bd7-9e42-ab0e89bfac90",
+      "target": "01f674f8-b3d1-4df1-acac-6cb8e0bfb63c",
+      "sourceHandle": "max_seq_len",
+      "targetHandle": "t5_max_seq_len"
+    },
+    {
+      "id": "reactflow__edge-f8d9d7c8-9ed7-4bd7-9e42-ab0e89bfac90t5_encoder-01f674f8-b3d1-4df1-acac-6cb8e0bfb63ct5_encoder",
+      "type": "default",
+      "source": "f8d9d7c8-9ed7-4bd7-9e42-ab0e89bfac90",
+      "target": "01f674f8-b3d1-4df1-acac-6cb8e0bfb63c",
+      "sourceHandle": "t5_encoder",
+      "targetHandle": "t5_encoder"
+    },
+    {
+      "id": "reactflow__edge-f8d9d7c8-9ed7-4bd7-9e42-ab0e89bfac90clip-01f674f8-b3d1-4df1-acac-6cb8e0bfb63cclip",
+      "type": "default",
+      "source": "f8d9d7c8-9ed7-4bd7-9e42-ab0e89bfac90",
+      "target": "01f674f8-b3d1-4df1-acac-6cb8e0bfb63c",
+      "sourceHandle": "clip",
+      "targetHandle": "clip"
+    }
+  ]
+}

invokeai/app/services/workflow_records/default_workflows/Flux Text to Image.json

@@ -1,27 +1,35 @@
 {
   "name": "FLUX Text to Image",
   "author": "InvokeAI",
-  "description": "A simple text-to-image workflow using FLUX dev or schnell models. Prerequisite model downloads: T5 Encoder, CLIP-L Encoder, and FLUX VAE. Quantized and un-quantized versions can be found in the starter models tab within your Model Manager. We recommend 4 steps for FLUX schnell models and 30 steps for FLUX dev models.",
-  "version": "1.0.0",
+  "description": "A simple text-to-image workflow using FLUX dev or schnell models.",
+  "version": "1.0.4",
   "contact": "",
   "tags": "text2image, flux",
   "notes": "Prerequisite model downloads: T5 Encoder, CLIP-L Encoder, and FLUX VAE. Quantized and un-quantized versions can be found in the starter models tab within your Model Manager. We recommend 4 steps for FLUX schnell models and 30 steps for FLUX dev models.",
   "exposedFields": [
     {
-      "nodeId": "4f0207c2-ff40-41fd-b047-ad33fbb1c33a",
+      "nodeId": "f8d9d7c8-9ed7-4bd7-9e42-ab0e89bfac90",
       "fieldName": "model"
     },
+    {
+      "nodeId": "f8d9d7c8-9ed7-4bd7-9e42-ab0e89bfac90",
+      "fieldName": "t5_encoder_model"
+    },
+    {
+      "nodeId": "f8d9d7c8-9ed7-4bd7-9e42-ab0e89bfac90",
+      "fieldName": "clip_embed_model"
+    },
+    {
+      "nodeId": "f8d9d7c8-9ed7-4bd7-9e42-ab0e89bfac90",
+      "fieldName": "vae_model"
+    },
     {
       "nodeId": "01f674f8-b3d1-4df1-acac-6cb8e0bfb63c",
       "fieldName": "prompt"
     },
     {
-      "nodeId": "159bdf1b-79e7-4174-b86e-d40e646964c8",
+      "nodeId": "4fe24f07-f906-4f55-ab2c-9beee56ef5bd",
       "fieldName": "num_steps"
-    },
-    {
-      "nodeId": "4f0207c2-ff40-41fd-b047-ad33fbb1c33a",
-      "fieldName": "t5_encoder"
     }
   ],
   "meta": {
@@ -30,12 +38,127 @@
   },
   "nodes": [
     {
-      "id": "4f0207c2-ff40-41fd-b047-ad33fbb1c33a",
+      "id": "4fe24f07-f906-4f55-ab2c-9beee56ef5bd",
       "type": "invocation",
       "data": {
-        "id": "4f0207c2-ff40-41fd-b047-ad33fbb1c33a",
+        "id": "4fe24f07-f906-4f55-ab2c-9beee56ef5bd",
+        "type": "flux_denoise",
+        "version": "1.0.0",
+        "label": "",
+        "notes": "",
+        "isOpen": true,
+        "isIntermediate": true,
+        "useCache": true,
+        "inputs": {
+          "board": {
+            "name": "board",
+            "label": ""
+          },
+          "metadata": {
+            "name": "metadata",
+            "label": ""
+          },
+          "latents": {
+            "name": "latents",
+            "label": ""
+          },
+          "denoise_mask": {
+            "name": "denoise_mask",
+            "label": ""
+          },
+          "denoising_start": {
+            "name": "denoising_start",
+            "label": "",
+            "value": 0
+          },
+          "denoising_end": {
+            "name": "denoising_end",
+            "label": "",
+            "value": 1
+          },
+          "transformer": {
+            "name": "transformer",
+            "label": ""
+          },
+          "positive_text_conditioning": {
+            "name": "positive_text_conditioning",
+            "label": ""
+          },
+          "width": {
+            "name": "width",
+            "label": "",
+            "value": 1024
+          },
+          "height": {
+            "name": "height",
+            "label": "",
+            "value": 1024
+          },
+          "num_steps": {
+            "name": "num_steps",
+            "label": "Steps (Recommend 30 for Dev, 4 for Schnell)",
+            "value": 30
+          },
+          "guidance": {
+            "name": "guidance",
+            "label": "",
+            "value": 4
+          },
+          "seed": {
+            "name": "seed",
+            "label": "",
+            "value": 0
+          }
+        }
+      },
+      "position": {
+        "x": 1186.1868226120378,
+        "y": -214.9459927686657
+      }
+    },
+    {
+      "id": "7e5172eb-48c1-44db-a770-8fd83e1435d1",
+      "type": "invocation",
+      "data": {
+        "id": "7e5172eb-48c1-44db-a770-8fd83e1435d1",
+        "type": "flux_vae_decode",
+        "version": "1.0.0",
+        "label": "",
+        "notes": "",
+        "isOpen": true,
+        "isIntermediate": false,
+        "useCache": true,
+        "inputs": {
+          "board": {
+            "name": "board",
+            "label": ""
+          },
+          "metadata": {
+            "name": "metadata",
+            "label": ""
+          },
+          "latents": {
+            "name": "latents",
+            "label": ""
+          },
+          "vae": {
+            "name": "vae",
+            "label": ""
+          }
+        }
+      },
+      "position": {
+        "x": 1575.5797431839133,
+        "y": -209.00150975507415
+      }
+    },
+    {
+      "id": "f8d9d7c8-9ed7-4bd7-9e42-ab0e89bfac90",
+      "type": "invocation",
+      "data": {
+        "id": "f8d9d7c8-9ed7-4bd7-9e42-ab0e89bfac90",
         "type": "flux_model_loader",
-        "version": "1.0.3",
+        "version": "1.0.4",
         "label": "",
         "notes": "",
         "isOpen": true,
@@ -44,31 +167,25 @@
         "inputs": {
           "model": {
             "name": "model",
-            "label": "Model (Starter Models can be found in Model Manager)",
-            "value": {
-              "key": "f04a7a2f-c74d-4538-8d5e-879a53501662",
-              "hash": "random:4875da7a9508444ffa706f61961c260d0c6729f6181a86b31fad06df1277b850",
-              "name": "FLUX Dev (Quantized)",
-              "base": "flux",
-              "type": "main"
-            }
+            "label": ""
           },
-          "t5_encoder": {
-            "name": "t5_encoder",
-            "label": "T 5 Encoder (Starter Models can be found in Model Manager)",
-            "value": {
-              "key": "20dcd9ec-5fbb-4012-8401-049e707da5e5",
-              "hash": "random:f986be43ff3502169e4adbdcee158afb0e0a65a1edc4cab16ae59963630cfd8f",
-              "name": "t5_bnb_int8_quantized_encoder",
-              "base": "any",
-              "type": "t5_encoder"
-            }
+          "t5_encoder_model": {
+            "name": "t5_encoder_model",
+            "label": ""
+          },
+          "clip_embed_model": {
+            "name": "clip_embed_model",
+            "label": ""
+          },
+          "vae_model": {
+            "name": "vae_model",
+            "label": ""
           }
         }
       },
       "position": {
-        "x": 337.09365228062825,
-        "y": 40.63469521079861
+        "x": 381.1882713063478,
+        "y": -95.89663532854017
       }
     },
     {
@@ -105,8 +222,8 @@
         }
       },
       "position": {
-        "x": 824.1970602278849,
-        "y": 146.98251001061735
+        "x": 778.4899149328337,
+        "y": -100.36469216659502
       }
     },
     {
@@ -135,132 +252,75 @@
         }
       },
       "position": {
-        "x": 822.9899179655476,
-        "y": 360.9657214885052
-      }
-    },
-    {
-      "id": "159bdf1b-79e7-4174-b86e-d40e646964c8",
-      "type": "invocation",
-      "data": {
-        "id": "159bdf1b-79e7-4174-b86e-d40e646964c8",
-        "type": "flux_text_to_image",
-        "version": "1.0.0",
-        "label": "",
-        "notes": "",
-        "isOpen": true,
-        "isIntermediate": false,
-        "useCache": true,
-        "inputs": {
-          "board": {
-            "name": "board",
-            "label": ""
-          },
-          "metadata": {
-            "name": "metadata",
-            "label": ""
-          },
-          "transformer": {
-            "name": "transformer",
-            "label": ""
-          },
-          "vae": {
-            "name": "vae",
-            "label": ""
-          },
-          "positive_text_conditioning": {
-            "name": "positive_text_conditioning",
-            "label": ""
-          },
-          "width": {
-            "name": "width",
-            "label": "",
-            "value": 1024
-          },
-          "height": {
-            "name": "height",
-            "label": "",
-            "value": 1024
-          },
-          "num_steps": {
-            "name": "num_steps",
-            "label": "Steps (Recommend 30 for Dev, 4 for Schnell)",
-            "value": 30
-          },
-          "guidance": {
-            "name": "guidance",
-            "label": "",
-            "value": 4
-          },
-          "seed": {
-            "name": "seed",
-            "label": "",
-            "value": 0
-          }
-        }
-      },
-      "position": {
-        "x": 1216.3900791301849,
-        "y": 5.500841807102248
+        "x": 800.9667463219505,
+        "y": 285.8297267547506
       }
     }
   ],
   "edges": [
     {
-      "id": "reactflow__edge-4f0207c2-ff40-41fd-b047-ad33fbb1c33amax_seq_len-01f674f8-b3d1-4df1-acac-6cb8e0bfb63ct5_max_seq_len",
+      "id": "reactflow__edge-f8d9d7c8-9ed7-4bd7-9e42-ab0e89bfac90transformer-4fe24f07-f906-4f55-ab2c-9beee56ef5bdtransformer",
       "type": "default",
-      "source": "4f0207c2-ff40-41fd-b047-ad33fbb1c33a",
+      "source": "f8d9d7c8-9ed7-4bd7-9e42-ab0e89bfac90",
+      "target": "4fe24f07-f906-4f55-ab2c-9beee56ef5bd",
+      "sourceHandle": "transformer",
+      "targetHandle": "transformer"
+    },
+    {
+      "id": "reactflow__edge-01f674f8-b3d1-4df1-acac-6cb8e0bfb63cconditioning-4fe24f07-f906-4f55-ab2c-9beee56ef5bdpositive_text_conditioning",
+      "type": "default",
+      "source": "01f674f8-b3d1-4df1-acac-6cb8e0bfb63c",
+      "target": "4fe24f07-f906-4f55-ab2c-9beee56ef5bd",
+      "sourceHandle": "conditioning",
+      "targetHandle": "positive_text_conditioning"
+    },
+    {
+      "id": "reactflow__edge-4754c534-a5f3-4ad0-9382-7887985e668cvalue-4fe24f07-f906-4f55-ab2c-9beee56ef5bdseed",
+      "type": "default",
+      "source": "4754c534-a5f3-4ad0-9382-7887985e668c",
+      "target": "4fe24f07-f906-4f55-ab2c-9beee56ef5bd",
+      "sourceHandle": "value",
+      "targetHandle": "seed"
+    },
+    {
+      "id": "reactflow__edge-4fe24f07-f906-4f55-ab2c-9beee56ef5bdlatents-7e5172eb-48c1-44db-a770-8fd83e1435d1latents",
+      "type": "default",
+      "source": "4fe24f07-f906-4f55-ab2c-9beee56ef5bd",
+      "target": "7e5172eb-48c1-44db-a770-8fd83e1435d1",
+      "sourceHandle": "latents",
+      "targetHandle": "latents"
+    },
+    {
+      "id": "reactflow__edge-f8d9d7c8-9ed7-4bd7-9e42-ab0e89bfac90vae-7e5172eb-48c1-44db-a770-8fd83e1435d1vae",
+      "type": "default",
+      "source": "f8d9d7c8-9ed7-4bd7-9e42-ab0e89bfac90",
+      "target": "7e5172eb-48c1-44db-a770-8fd83e1435d1",
+      "sourceHandle": "vae",
+      "targetHandle": "vae"
+    },
+    {
+      "id": "reactflow__edge-f8d9d7c8-9ed7-4bd7-9e42-ab0e89bfac90max_seq_len-01f674f8-b3d1-4df1-acac-6cb8e0bfb63ct5_max_seq_len",
+      "type": "default",
+      "source": "f8d9d7c8-9ed7-4bd7-9e42-ab0e89bfac90",
       "target": "01f674f8-b3d1-4df1-acac-6cb8e0bfb63c",
       "sourceHandle": "max_seq_len",
       "targetHandle": "t5_max_seq_len"
     },
     {
-      "id": "reactflow__edge-4f0207c2-ff40-41fd-b047-ad33fbb1c33avae-159bdf1b-79e7-4174-b86e-d40e646964c8vae",
+      "id": "reactflow__edge-f8d9d7c8-9ed7-4bd7-9e42-ab0e89bfac90t5_encoder-01f674f8-b3d1-4df1-acac-6cb8e0bfb63ct5_encoder",
       "type": "default",
-      "source": "4f0207c2-ff40-41fd-b047-ad33fbb1c33a",
-      "target": "159bdf1b-79e7-4174-b86e-d40e646964c8",
-      "sourceHandle": "vae",
-      "targetHandle": "vae"
-    },
-    {
-      "id": "reactflow__edge-4f0207c2-ff40-41fd-b047-ad33fbb1c33atransformer-159bdf1b-79e7-4174-b86e-d40e646964c8transformer",
-      "type": "default",
-      "source": "4f0207c2-ff40-41fd-b047-ad33fbb1c33a",
-      "target": "159bdf1b-79e7-4174-b86e-d40e646964c8",
-      "sourceHandle": "transformer",
-      "targetHandle": "transformer"
-    },
-    {
-      "id": "reactflow__edge-4f0207c2-ff40-41fd-b047-ad33fbb1c33at5_encoder-01f674f8-b3d1-4df1-acac-6cb8e0bfb63ct5_encoder",
-      "type": "default",
-      "source": "4f0207c2-ff40-41fd-b047-ad33fbb1c33a",
+      "source": "f8d9d7c8-9ed7-4bd7-9e42-ab0e89bfac90",
       "target": "01f674f8-b3d1-4df1-acac-6cb8e0bfb63c",
       "sourceHandle": "t5_encoder",
       "targetHandle": "t5_encoder"
     },
     {
-      "id": "reactflow__edge-4f0207c2-ff40-41fd-b047-ad33fbb1c33aclip-01f674f8-b3d1-4df1-acac-6cb8e0bfb63cclip",
+      "id": "reactflow__edge-f8d9d7c8-9ed7-4bd7-9e42-ab0e89bfac90clip-01f674f8-b3d1-4df1-acac-6cb8e0bfb63cclip",
       "type": "default",
-      "source": "4f0207c2-ff40-41fd-b047-ad33fbb1c33a",
+      "source": "f8d9d7c8-9ed7-4bd7-9e42-ab0e89bfac90",
       "target": "01f674f8-b3d1-4df1-acac-6cb8e0bfb63c",
       "sourceHandle": "clip",
       "targetHandle": "clip"
-    },
-    {
-      "id": "reactflow__edge-01f674f8-b3d1-4df1-acac-6cb8e0bfb63cconditioning-159bdf1b-79e7-4174-b86e-d40e646964c8positive_text_conditioning",
-      "type": "default",
-      "source": "01f674f8-b3d1-4df1-acac-6cb8e0bfb63c",
-      "target": "159bdf1b-79e7-4174-b86e-d40e646964c8",
-      "sourceHandle": "conditioning",
-      "targetHandle": "positive_text_conditioning"
-    },
-    {
-      "id": "reactflow__edge-4754c534-a5f3-4ad0-9382-7887985e668cvalue-159bdf1b-79e7-4174-b86e-d40e646964c8seed",
-      "type": "default",
-      "source": "4754c534-a5f3-4ad0-9382-7887985e668c",
-      "target": "159bdf1b-79e7-4174-b86e-d40e646964c8",
-      "sourceHandle": "value",
-      "targetHandle": "seed"
     }
   ]
 }

invokeai/app/util/step_callback.py

@@ -38,6 +38,25 @@ SD1_5_LATENT_RGB_FACTORS = [
     [-0.1307, -0.1874, -0.7445],  # L4
 ]
 
+FLUX_LATENT_RGB_FACTORS = [
+    [-0.0412, 0.0149, 0.0521],
+    [0.0056, 0.0291, 0.0768],
+    [0.0342, -0.0681, -0.0427],
+    [-0.0258, 0.0092, 0.0463],
+    [0.0863, 0.0784, 0.0547],
+    [-0.0017, 0.0402, 0.0158],
+    [0.0501, 0.1058, 0.1152],
+    [-0.0209, -0.0218, -0.0329],
+    [-0.0314, 0.0083, 0.0896],
+    [0.0851, 0.0665, -0.0472],
+    [-0.0534, 0.0238, -0.0024],
+    [0.0452, -0.0026, 0.0048],
+    [0.0892, 0.0831, 0.0881],
+    [-0.1117, -0.0304, -0.0789],
+    [0.0027, -0.0479, -0.0043],
+    [-0.1146, -0.0827, -0.0598],
+]
+
 
 def sample_to_lowres_estimated_image(
     samples: torch.Tensor, latent_rgb_factors: torch.Tensor, smooth_matrix: Optional[torch.Tensor] = None
@@ -94,3 +113,32 @@ def stable_diffusion_step_callback(
         intermediate_state,
         ProgressImage(dataURL=dataURL, width=width, height=height),
     )
+
+
+def flux_step_callback(
+    context_data: "InvocationContextData",
+    intermediate_state: PipelineIntermediateState,
+    events: "EventServiceBase",
+    is_canceled: Callable[[], bool],
+) -> None:
+    if is_canceled():
+        raise CanceledException
+    sample = intermediate_state.latents
+    latent_rgb_factors = torch.tensor(FLUX_LATENT_RGB_FACTORS, dtype=sample.dtype, device=sample.device)
+    latent_image_perm = sample.permute(1, 2, 0).to(dtype=sample.dtype, device=sample.device)
+    latent_image = latent_image_perm @ latent_rgb_factors
+    latents_ubyte = (
+        ((latent_image + 1) / 2).clamp(0, 1).mul(0xFF)  # change scale from -1..1 to 0..1  # to 0..255
+    ).to(device="cpu", dtype=torch.uint8)
+    image = Image.fromarray(latents_ubyte.cpu().numpy())
+    (width, height) = image.size
+    width *= 8
+    height *= 8
+    dataURL = image_to_dataURL(image, image_format="JPEG")
+
+    events.emit_invocation_denoise_progress(
+        context_data.queue_item,
+        context_data.invocation,
+        intermediate_state,
+        ProgressImage(dataURL=dataURL, width=width, height=height),
+    )

invokeai/backend/flux/denoise.py

@@ -0,0 +1,56 @@
+from typing import Callable
+
+import torch
+from tqdm import tqdm
+
+from invokeai.backend.flux.inpaint_extension import InpaintExtension
+from invokeai.backend.flux.model import Flux
+from invokeai.backend.stable_diffusion.diffusers_pipeline import PipelineIntermediateState
+
+
+def denoise(
+    model: Flux,
+    # model input
+    img: torch.Tensor,
+    img_ids: torch.Tensor,
+    txt: torch.Tensor,
+    txt_ids: torch.Tensor,
+    vec: torch.Tensor,
+    # sampling parameters
+    timesteps: list[float],
+    step_callback: Callable[[PipelineIntermediateState], None],
+    guidance: float,
+    inpaint_extension: InpaintExtension | None,
+):
+    step = 0
+    # guidance_vec is ignored for schnell.
+    guidance_vec = torch.full((img.shape[0],), guidance, device=img.device, dtype=img.dtype)
+    for t_curr, t_prev in tqdm(list(zip(timesteps[:-1], timesteps[1:], strict=True))):
+        t_vec = torch.full((img.shape[0],), t_curr, dtype=img.dtype, device=img.device)
+        pred = model(
+            img=img,
+            img_ids=img_ids,
+            txt=txt,
+            txt_ids=txt_ids,
+            y=vec,
+            timesteps=t_vec,
+            guidance=guidance_vec,
+        )
+        preview_img = img - t_curr * pred
+        img = img + (t_prev - t_curr) * pred
+
+        if inpaint_extension is not None:
+            img = inpaint_extension.merge_intermediate_latents_with_init_latents(img, t_prev)
+
+        step_callback(
+            PipelineIntermediateState(
+                step=step,
+                order=1,
+                total_steps=len(timesteps),
+                timestep=int(t_curr),
+                latents=preview_img,
+            ),
+        )
+        step += 1
+
+    return img

invokeai/backend/flux/inpaint_extension.py

@@ -0,0 +1,35 @@
+import torch
+
+
+class InpaintExtension:
+    """A class for managing inpainting with FLUX."""
+
+    def __init__(self, init_latents: torch.Tensor, inpaint_mask: torch.Tensor, noise: torch.Tensor):
+        """Initialize InpaintExtension.
+
+        Args:
+            init_latents (torch.Tensor): The initial latents (i.e. un-noised at timestep 0). In 'packed' format.
+            inpaint_mask (torch.Tensor): A mask specifying which elements to inpaint. Range [0, 1]. Values of 1 will be
+                re-generated. Values of 0 will remain unchanged. Values between 0 and 1 can be used to blend the
+                inpainted region with the background. In 'packed' format.
+            noise (torch.Tensor): The noise tensor used to noise the init_latents. In 'packed' format.
+        """
+        assert init_latents.shape == inpaint_mask.shape == noise.shape
+        self._init_latents = init_latents
+        self._inpaint_mask = inpaint_mask
+        self._noise = noise
+
+    def merge_intermediate_latents_with_init_latents(
+        self, intermediate_latents: torch.Tensor, timestep: float
+    ) -> torch.Tensor:
+        """Merge the intermediate latents with the initial latents for the current timestep using the inpaint mask. I.e.
+        update the intermediate latents to keep the regions that are not being inpainted on the correct noise
+        trajectory.
+
+        This function should be called after each denoising step.
+        """
+        # Noise the init latents for the current timestep.
+        noised_init_latents = self._noise * timestep + (1.0 - timestep) * self._init_latents
+
+        # Merge the intermediate latents with the noised_init_latents using the inpaint_mask.
+        return intermediate_latents * self._inpaint_mask + noised_init_latents * (1.0 - self._inpaint_mask)

invokeai/backend/flux/modules/autoencoder.py

@@ -258,16 +258,17 @@ class Decoder(nn.Module):
 
 
 class DiagonalGaussian(nn.Module):
-    def __init__(self, sample: bool = True, chunk_dim: int = 1):
+    def __init__(self, chunk_dim: int = 1):
         super().__init__()
-        self.sample = sample
         self.chunk_dim = chunk_dim
 
-    def forward(self, z: Tensor) -> Tensor:
+    def forward(self, z: Tensor, sample: bool = True, generator: torch.Generator | None = None) -> Tensor:
         mean, logvar = torch.chunk(z, 2, dim=self.chunk_dim)
-        if self.sample:
+        if sample:
             std = torch.exp(0.5 * logvar)
-            return mean + std * torch.randn_like(mean)
+            # Unfortunately, torch.randn_like(...) does not accept a generator argument at the time of writing, so we
+            # have to use torch.randn(...) instead.
+            return mean + std * torch.randn(size=mean.size(), generator=generator, dtype=mean.dtype, device=mean.device)
         else:
             return mean
 
@@ -297,8 +298,21 @@ class AutoEncoder(nn.Module):
         self.scale_factor = params.scale_factor
         self.shift_factor = params.shift_factor
 
-    def encode(self, x: Tensor) -> Tensor:
-        z = self.reg(self.encoder(x))
+    def encode(self, x: Tensor, sample: bool = True, generator: torch.Generator | None = None) -> Tensor:
+        """Run VAE encoding on input tensor x.
+
+        Args:
+            x (Tensor): Input image tensor. Shape: (batch_size, in_channels, height, width).
+            sample (bool, optional): If True, sample from the encoded distribution, else, return the distribution mean.
+                Defaults to True.
+            generator (torch.Generator | None, optional): Optional random number generator for reproducibility.
+                Defaults to None.
+
+        Returns:
+            Tensor: Encoded latent tensor. Shape: (batch_size, z_channels, latent_height, latent_width).
+        """
+
+        z = self.reg(self.encoder(x), sample=sample, generator=generator)
         z = self.scale_factor * (z - self.shift_factor)
         return z
 

invokeai/backend/flux/sampling.py

@@ -1,176 +0,0 @@
-# Initially pulled from https://github.com/black-forest-labs/flux
-
-import math
-from typing import Callable
-
-import torch
-from einops import rearrange, repeat
-from torch import Tensor
-from tqdm import tqdm
-
-from invokeai.backend.flux.model import Flux
-from invokeai.backend.flux.modules.conditioner import HFEncoder
-
-
-def get_noise(
-    num_samples: int,
-    height: int,
-    width: int,
-    device: torch.device,
-    dtype: torch.dtype,
-    seed: int,
-):
-    # We always generate noise on the same device and dtype then cast to ensure consistency across devices/dtypes.
-    rand_device = "cpu"
-    rand_dtype = torch.float16
-    return torch.randn(
-        num_samples,
-        16,
-        # allow for packing
-        2 * math.ceil(height / 16),
-        2 * math.ceil(width / 16),
-        device=rand_device,
-        dtype=rand_dtype,
-        generator=torch.Generator(device=rand_device).manual_seed(seed),
-    ).to(device=device, dtype=dtype)
-
-
-def prepare(t5: HFEncoder, clip: HFEncoder, img: Tensor, prompt: str | list[str]) -> dict[str, Tensor]:
-    bs, c, h, w = img.shape
-    if bs == 1 and not isinstance(prompt, str):
-        bs = len(prompt)
-
-    img = rearrange(img, "b c (h ph) (w pw) -> b (h w) (c ph pw)", ph=2, pw=2)
-    if img.shape[0] == 1 and bs > 1:
-        img = repeat(img, "1 ... -> bs ...", bs=bs)
-
-    img_ids = torch.zeros(h // 2, w // 2, 3)
-    img_ids[..., 1] = img_ids[..., 1] + torch.arange(h // 2)[:, None]
-    img_ids[..., 2] = img_ids[..., 2] + torch.arange(w // 2)[None, :]
-    img_ids = repeat(img_ids, "h w c -> b (h w) c", b=bs)
-
-    if isinstance(prompt, str):
-        prompt = [prompt]
-    txt = t5(prompt)
-    if txt.shape[0] == 1 and bs > 1:
-        txt = repeat(txt, "1 ... -> bs ...", bs=bs)
-    txt_ids = torch.zeros(bs, txt.shape[1], 3)
-
-    vec = clip(prompt)
-    if vec.shape[0] == 1 and bs > 1:
-        vec = repeat(vec, "1 ... -> bs ...", bs=bs)
-
-    return {
-        "img": img,
-        "img_ids": img_ids.to(img.device),
-        "txt": txt.to(img.device),
-        "txt_ids": txt_ids.to(img.device),
-        "vec": vec.to(img.device),
-    }
-
-
-def time_shift(mu: float, sigma: float, t: Tensor):
-    return math.exp(mu) / (math.exp(mu) + (1 / t - 1) ** sigma)
-
-
-def get_lin_function(x1: float = 256, y1: float = 0.5, x2: float = 4096, y2: float = 1.15) -> Callable[[float], float]:
-    m = (y2 - y1) / (x2 - x1)
-    b = y1 - m * x1
-    return lambda x: m * x + b
-
-
-def get_schedule(
-    num_steps: int,
-    image_seq_len: int,
-    base_shift: float = 0.5,
-    max_shift: float = 1.15,
-    shift: bool = True,
-) -> list[float]:
-    # extra step for zero
-    timesteps = torch.linspace(1, 0, num_steps + 1)
-
-    # shifting the schedule to favor high timesteps for higher signal images
-    if shift:
-        # eastimate mu based on linear estimation between two points
-        mu = get_lin_function(y1=base_shift, y2=max_shift)(image_seq_len)
-        timesteps = time_shift(mu, 1.0, timesteps)
-
-    return timesteps.tolist()
-
-
-def denoise(
-    model: Flux,
-    # model input
-    img: Tensor,
-    img_ids: Tensor,
-    txt: Tensor,
-    txt_ids: Tensor,
-    vec: Tensor,
-    # sampling parameters
-    timesteps: list[float],
-    step_callback: Callable[[], None],
-    guidance: float = 4.0,
-):
-    dtype = model.txt_in.bias.dtype
-
-    # TODO(ryand): This shouldn't be necessary if we manage the dtypes properly in the caller.
-    img = img.to(dtype=dtype)
-    img_ids = img_ids.to(dtype=dtype)
-    txt = txt.to(dtype=dtype)
-    txt_ids = txt_ids.to(dtype=dtype)
-    vec = vec.to(dtype=dtype)
-
-    # this is ignored for schnell
-    guidance_vec = torch.full((img.shape[0],), guidance, device=img.device, dtype=img.dtype)
-    for t_curr, t_prev in tqdm(list(zip(timesteps[:-1], timesteps[1:], strict=True))):
-        t_vec = torch.full((img.shape[0],), t_curr, dtype=img.dtype, device=img.device)
-        pred = model(
-            img=img,
-            img_ids=img_ids,
-            txt=txt,
-            txt_ids=txt_ids,
-            y=vec,
-            timesteps=t_vec,
-            guidance=guidance_vec,
-        )
-
-        img = img + (t_prev - t_curr) * pred
-        step_callback()
-
-    return img
-
-
-def unpack(x: Tensor, height: int, width: int) -> Tensor:
-    return rearrange(
-        x,
-        "b (h w) (c ph pw) -> b c (h ph) (w pw)",
-        h=math.ceil(height / 16),
-        w=math.ceil(width / 16),
-        ph=2,
-        pw=2,
-    )
-
-
-def prepare_latent_img_patches(latent_img: torch.Tensor) -> tuple[torch.Tensor, torch.Tensor]:
-    """Convert an input image in latent space to patches for diffusion.
-
-    This implementation was extracted from:
-    https://github.com/black-forest-labs/flux/blob/c00d7c60b085fce8058b9df845e036090873f2ce/src/flux/sampling.py#L32
-
-    Returns:
-        tuple[Tensor, Tensor]: (img, img_ids), as defined in the original flux repo.
-    """
-    bs, c, h, w = latent_img.shape
-
-    # Pixel unshuffle with a scale of 2, and flatten the height/width dimensions to get an array of patches.
-    img = rearrange(latent_img, "b c (h ph) (w pw) -> b (h w) (c ph pw)", ph=2, pw=2)
-    if img.shape[0] == 1 and bs > 1:
-        img = repeat(img, "1 ... -> bs ...", bs=bs)
-
-    # Generate patch position ids.
-    img_ids = torch.zeros(h // 2, w // 2, 3, device=img.device)
-    img_ids[..., 1] = img_ids[..., 1] + torch.arange(h // 2, device=img.device)[:, None]
-    img_ids[..., 2] = img_ids[..., 2] + torch.arange(w // 2, device=img.device)[None, :]
-    img_ids = repeat(img_ids, "h w c -> b (h w) c", b=bs)
-
-    return img, img_ids

invokeai/backend/flux/sampling_utils.py

@@ -0,0 +1,135 @@
+# Initially pulled from https://github.com/black-forest-labs/flux
+
+import math
+from typing import Callable
+
+import torch
+from einops import rearrange, repeat
+
+
+def get_noise(
+    num_samples: int,
+    height: int,
+    width: int,
+    device: torch.device,
+    dtype: torch.dtype,
+    seed: int,
+):
+    # We always generate noise on the same device and dtype then cast to ensure consistency across devices/dtypes.
+    rand_device = "cpu"
+    rand_dtype = torch.float16
+    return torch.randn(
+        num_samples,
+        16,
+        # allow for packing
+        2 * math.ceil(height / 16),
+        2 * math.ceil(width / 16),
+        device=rand_device,
+        dtype=rand_dtype,
+        generator=torch.Generator(device=rand_device).manual_seed(seed),
+    ).to(device=device, dtype=dtype)
+
+
+def time_shift(mu: float, sigma: float, t: torch.Tensor) -> torch.Tensor:
+    return math.exp(mu) / (math.exp(mu) + (1 / t - 1) ** sigma)
+
+
+def get_lin_function(x1: float = 256, y1: float = 0.5, x2: float = 4096, y2: float = 1.15) -> Callable[[float], float]:
+    m = (y2 - y1) / (x2 - x1)
+    b = y1 - m * x1
+    return lambda x: m * x + b
+
+
+def get_schedule(
+    num_steps: int,
+    image_seq_len: int,
+    base_shift: float = 0.5,
+    max_shift: float = 1.15,
+    shift: bool = True,
+) -> list[float]:
+    # extra step for zero
+    timesteps = torch.linspace(1, 0, num_steps + 1)
+
+    # shifting the schedule to favor high timesteps for higher signal images
+    if shift:
+        # estimate mu based on linear estimation between two points
+        mu = get_lin_function(y1=base_shift, y2=max_shift)(image_seq_len)
+        timesteps = time_shift(mu, 1.0, timesteps)
+
+    return timesteps.tolist()
+
+
+def _find_last_index_ge_val(timesteps: list[float], val: float, eps: float = 1e-6) -> int:
+    """Find the last index in timesteps that is >= val.
+
+    We use epsilon-close equality to avoid potential floating point errors.
+    """
+    idx = len(list(filter(lambda t: t >= (val - eps), timesteps))) - 1
+    assert idx >= 0
+    return idx
+
+
+def clip_timestep_schedule(timesteps: list[float], denoising_start: float, denoising_end: float) -> list[float]:
+    """Clip the timestep schedule to the denoising range.
+
+    Args:
+        timesteps (list[float]): The original timestep schedule: [1.0, ..., 0.0].
+        denoising_start (float): A value in [0, 1] specifying the start of the denoising process. E.g. a value of 0.2
+            would mean that the denoising process start at the last timestep in the schedule >= 0.8.
+        denoising_end (float): A value in [0, 1] specifying the end of the denoising process. E.g. a value of 0.8 would
+            mean that the denoising process end at the last timestep in the schedule >= 0.2.
+
+    Returns:
+        list[float]: The clipped timestep schedule.
+    """
+    assert 0.0 <= denoising_start <= 1.0
+    assert 0.0 <= denoising_end <= 1.0
+    assert denoising_start <= denoising_end
+
+    t_start_val = 1.0 - denoising_start
+    t_end_val = 1.0 - denoising_end
+
+    t_start_idx = _find_last_index_ge_val(timesteps, t_start_val)
+    t_end_idx = _find_last_index_ge_val(timesteps, t_end_val)
+
+    clipped_timesteps = timesteps[t_start_idx : t_end_idx + 1]
+
+    return clipped_timesteps
+
+
+def unpack(x: torch.Tensor, height: int, width: int) -> torch.Tensor:
+    """Unpack flat array of patch embeddings to latent image."""
+    return rearrange(
+        x,
+        "b (h w) (c ph pw) -> b c (h ph) (w pw)",
+        h=math.ceil(height / 16),
+        w=math.ceil(width / 16),
+        ph=2,
+        pw=2,
+    )
+
+
+def pack(x: torch.Tensor) -> torch.Tensor:
+    """Pack latent image to flattented array of patch embeddings."""
+    # Pixel unshuffle with a scale of 2, and flatten the height/width dimensions to get an array of patches.
+    return rearrange(x, "b c (h ph) (w pw) -> b (h w) (c ph pw)", ph=2, pw=2)
+
+
+def generate_img_ids(h: int, w: int, batch_size: int, device: torch.device, dtype: torch.dtype) -> torch.Tensor:
+    """Generate tensor of image position ids.
+
+    Args:
+        h (int): Height of image in latent space.
+        w (int): Width of image in latent space.
+        batch_size (int): Batch size.
+        device (torch.device): Device.
+        dtype (torch.dtype): dtype.
+
+    Returns:
+        torch.Tensor: Image position ids.
+    """
+    img_ids = torch.zeros(h // 2, w // 2, 3, device=device, dtype=dtype)
+    img_ids[..., 1] = img_ids[..., 1] + torch.arange(h // 2, device=device, dtype=dtype)[:, None]
+    img_ids[..., 2] = img_ids[..., 2] + torch.arange(w // 2, device=device, dtype=dtype)[None, :]
+    img_ids = repeat(img_ids, "h w c -> b (h w) c", b=batch_size)
+    return img_ids

invokeai/backend/model_manager/load/load_default.py

@@ -66,12 +66,14 @@ class ModelLoader(ModelLoaderBase):
         return (model_base / config.path).resolve()
 
     def _load_and_cache(self, config: AnyModelConfig, submodel_type: Optional[SubModelType] = None) -> ModelLockerBase:
+        stats_name = ":".join([config.base, config.type, config.name, (submodel_type or "")])
         try:
-            return self._ram_cache.get(config.key, submodel_type)
+            return self._ram_cache.get(config.key, submodel_type, stats_name=stats_name)
         except IndexError:
             pass
 
         config.path = str(self._get_model_path(config))
+        self._ram_cache.make_room(self.get_size_fs(config, Path(config.path), submodel_type))
         loaded_model = self._load_model(config, submodel_type)
 
         self._ram_cache.put(
@@ -83,7 +85,7 @@ class ModelLoader(ModelLoaderBase):
         return self._ram_cache.get(
             key=config.key,
             submodel_type=submodel_type,
-            stats_name=":".join([config.base, config.type, config.name, (submodel_type or "")]),
+            stats_name=stats_name,
         )
 
     def get_size_fs(

invokeai/backend/model_manager/load/model_cache/model_cache_base.py

@@ -128,7 +128,24 @@ class ModelCacheBase(ABC, Generic[T]):
     @property
     @abstractmethod
     def max_cache_size(self) -> float:
-        """Return true if the cache is configured to lazily offload models in VRAM."""
+        """Return the maximum size the RAM cache can grow to."""
+        pass
+
+    @max_cache_size.setter
+    @abstractmethod
+    def max_cache_size(self, value: float) -> None:
+        """Set the cap on vram cache size."""
+
+    @property
+    @abstractmethod
+    def max_vram_cache_size(self) -> float:
+        """Return the maximum size the VRAM cache can grow to."""
+        pass
+
+    @max_vram_cache_size.setter
+    @abstractmethod
+    def max_vram_cache_size(self, value: float) -> float:
+        """Set the maximum size the VRAM cache can grow to."""
         pass
 
     @abstractmethod
@@ -193,15 +210,6 @@ class ModelCacheBase(ABC, Generic[T]):
         """
         pass
 
-    @abstractmethod
-    def exists(
-        self,
-        key: str,
-        submodel_type: Optional[SubModelType] = None,
-    ) -> bool:
-        """Return true if the model identified by key and submodel_type is in the cache."""
-        pass
-
     @abstractmethod
     def cache_size(self) -> int:
         """Get the total size of the models currently cached."""

invokeai/backend/model_manager/load/model_cache/model_cache_default.py

@@ -1,22 +1,6 @@
 # Copyright (c) 2024 Lincoln D. Stein and the InvokeAI Development team
 # TODO: Add Stalker's proper name to copyright
-"""
-Manage a RAM cache of diffusion/transformer models for fast switching.
-They are moved between GPU VRAM and CPU RAM as necessary. If the cache
-grows larger than a preset maximum, then the least recently used
-model will be cleared and (re)loaded from disk when next needed.
-
-The cache returns context manager generators designed to load the
-model into the GPU within the context, and unload outside the
-context. Use like this:
-
-   cache = ModelCache(max_cache_size=7.5)
-   with cache.get_model('runwayml/stable-diffusion-1-5') as SD1,
-          cache.get_model('stabilityai/stable-diffusion-2') as SD2:
-       do_something_in_GPU(SD1,SD2)
-
-
-"""
+""" """
 
 import gc
 import math
@@ -40,53 +24,74 @@ from invokeai.backend.model_manager.load.model_util import calc_model_size_by_da
 from invokeai.backend.util.devices import TorchDevice
 from invokeai.backend.util.logging import InvokeAILogger
 
-# Maximum size of the cache, in gigs
-# Default is roughly enough to hold three fp16 diffusers models in RAM simultaneously
-DEFAULT_MAX_CACHE_SIZE = 6.0
-
-# amount of GPU memory to hold in reserve for use by generations (GB)
-DEFAULT_MAX_VRAM_CACHE_SIZE = 2.75
-
-# actual size of a gig
-GIG = 1073741824
+# Size of a GB in bytes.
+GB = 2**30
 
 # Size of a MB in bytes.
 MB = 2**20
 
 
 class ModelCache(ModelCacheBase[AnyModel]):
-    """Implementation of ModelCacheBase."""
+    """A cache for managing models in memory.
+
+    The cache is based on two levels of model storage:
+    - execution_device: The device where most models are executed (typically "cuda", "mps", or "cpu").
+    - storage_device: The device where models are offloaded when not in active use (typically "cpu").
+
+    The model cache is based on the following assumptions:
+    - storage_device_mem_size > execution_device_mem_size
+    - disk_to_storage_device_transfer_time >> storage_device_to_execution_device_transfer_time
+
+    A copy of all models in the cache is always kept on the storage_device. A subset of the models also have a copy on
+    the execution_device.
+
+    Models are moved between the storage_device and the execution_device as necessary. Cache size limits are enforced
+    on both the storage_device and the execution_device. The execution_device cache uses a smallest-first offload
+    policy. The storage_device cache uses a least-recently-used (LRU) offload policy.
+
+    Note: Neither of these offload policies has really been compared against alternatives. It's likely that different
+    policies would be better, although the optimal policies are likely heavily dependent on usage patterns and HW
+    configuration.
+
+    The cache returns context manager generators designed to load the model into the execution device (often GPU) within
+    the context, and unload outside the context.
+
+    Example usage:
+    ```
+    cache = ModelCache(max_cache_size=7.5, max_vram_cache_size=6.0)
+    with cache.get_model('runwayml/stable-diffusion-1-5') as SD1:
+        do_something_on_gpu(SD1)
+    ```
+    """
 
     def __init__(
         self,
-        max_cache_size: float = DEFAULT_MAX_CACHE_SIZE,
-        max_vram_cache_size: float = DEFAULT_MAX_VRAM_CACHE_SIZE,
+        max_cache_size: float,
+        max_vram_cache_size: float,
         execution_device: torch.device = torch.device("cuda"),
         storage_device: torch.device = torch.device("cpu"),
         precision: torch.dtype = torch.float16,
-        sequential_offload: bool = False,
         lazy_offloading: bool = True,
-        sha_chunksize: int = 16777216,
         log_memory_usage: bool = False,
         logger: Optional[Logger] = None,
     ):
         """
         Initialize the model RAM cache.
 
-        :param max_cache_size: Maximum size of the RAM cache [6.0 GB]
+        :param max_cache_size: Maximum size of the storage_device cache in GBs.
+        :param max_vram_cache_size: Maximum size of the execution_device cache in GBs.
         :param execution_device: Torch device to load active model into [torch.device('cuda')]
         :param storage_device: Torch device to save inactive model in [torch.device('cpu')]
         :param precision: Precision for loaded models [torch.float16]
         :param lazy_offloading: Keep model in VRAM until another model needs to be loaded
-        :param sequential_offload: Conserve VRAM by loading and unloading each stage of the pipeline sequentially
         :param log_memory_usage: If True, a memory snapshot will be captured before and after every model cache
             operation, and the result will be logged (at debug level). There is a time cost to capturing the memory
             snapshots, so it is recommended to disable this feature unless you are actively inspecting the model cache's
             behaviour.
+        :param logger: InvokeAILogger to use (otherwise creates one)
         """
         # allow lazy offloading only when vram cache enabled
         self._lazy_offloading = lazy_offloading and max_vram_cache_size > 0
-        self._precision: torch.dtype = precision
         self._max_cache_size: float = max_cache_size
         self._max_vram_cache_size: float = max_vram_cache_size
         self._execution_device: torch.device = execution_device
@@ -128,6 +133,16 @@ class ModelCache(ModelCacheBase[AnyModel]):
         """Set the cap on cache size."""
         self._max_cache_size = value
 
+    @property
+    def max_vram_cache_size(self) -> float:
+        """Return the cap on vram cache size."""
+        return self._max_vram_cache_size
+
+    @max_vram_cache_size.setter
+    def max_vram_cache_size(self, value: float) -> None:
+        """Set the cap on vram cache size."""
+        self._max_vram_cache_size = value
+
     @property
     def stats(self) -> Optional[CacheStats]:
         """Return collected CacheStats object."""
@@ -145,15 +160,6 @@ class ModelCache(ModelCacheBase[AnyModel]):
             total += cache_record.size
         return total
 
-    def exists(
-        self,
-        key: str,
-        submodel_type: Optional[SubModelType] = None,
-    ) -> bool:
-        """Return true if the model identified by key and submodel_type is in the cache."""
-        key = self._make_cache_key(key, submodel_type)
-        return key in self._cached_models
-
     def put(
         self,
         key: str,
@@ -203,7 +209,7 @@ class ModelCache(ModelCacheBase[AnyModel]):
         # more stats
         if self.stats:
             stats_name = stats_name or key
-            self.stats.cache_size = int(self._max_cache_size * GIG)
+            self.stats.cache_size = int(self._max_cache_size * GB)
             self.stats.high_watermark = max(self.stats.high_watermark, self.cache_size())
             self.stats.in_cache = len(self._cached_models)
             self.stats.loaded_model_sizes[stats_name] = max(
@@ -231,10 +237,13 @@ class ModelCache(ModelCacheBase[AnyModel]):
             return model_key
 
     def offload_unlocked_models(self, size_required: int) -> None:
-        """Move any unused models from VRAM."""
-        reserved = self._max_vram_cache_size * GIG
+        """Offload models from the execution_device to make room for size_required.
+
+        :param size_required: The amount of space to clear in the execution_device cache, in bytes.
+        """
+        reserved = self._max_vram_cache_size * GB
         vram_in_use = torch.cuda.memory_allocated() + size_required
-        self.logger.debug(f"{(vram_in_use/GIG):.2f}GB VRAM needed for models; max allowed={(reserved/GIG):.2f}GB")
+        self.logger.debug(f"{(vram_in_use/GB):.2f}GB VRAM needed for models; max allowed={(reserved/GB):.2f}GB")
         for _, cache_entry in sorted(self._cached_models.items(), key=lambda x: x[1].size):
             if vram_in_use <= reserved:
                 break
@@ -245,7 +254,7 @@ class ModelCache(ModelCacheBase[AnyModel]):
                 cache_entry.loaded = False
                 vram_in_use = torch.cuda.memory_allocated() + size_required
                 self.logger.debug(
-                    f"Removing {cache_entry.key} from VRAM to free {(cache_entry.size/GIG):.2f}GB; vram free = {(torch.cuda.memory_allocated()/GIG):.2f}GB"
+                    f"Removing {cache_entry.key} from VRAM to free {(cache_entry.size/GB):.2f}GB; vram free = {(torch.cuda.memory_allocated()/GB):.2f}GB"
                 )
 
         TorchDevice.empty_cache()
@@ -303,7 +312,7 @@ class ModelCache(ModelCacheBase[AnyModel]):
         self.logger.debug(
             f"Moved model '{cache_entry.key}' from {source_device} to"
             f" {target_device} in {(end_model_to_time-start_model_to_time):.2f}s."
-            f"Estimated model size: {(cache_entry.size/GIG):.3f} GB."
+            f"Estimated model size: {(cache_entry.size/GB):.3f} GB."
             f"{get_pretty_snapshot_diff(snapshot_before, snapshot_after)}"
         )
 
@@ -326,14 +335,14 @@ class ModelCache(ModelCacheBase[AnyModel]):
                     f"Moving model '{cache_entry.key}' from {source_device} to"
                     f" {target_device} caused an unexpected change in VRAM usage. The model's"
                     " estimated size may be incorrect. Estimated model size:"
-                    f" {(cache_entry.size/GIG):.3f} GB.\n"
+                    f" {(cache_entry.size/GB):.3f} GB.\n"
                     f"{get_pretty_snapshot_diff(snapshot_before, snapshot_after)}"
                 )
 
     def print_cuda_stats(self) -> None:
         """Log CUDA diagnostics."""
-        vram = "%4.2fG" % (torch.cuda.memory_allocated() / GIG)
-        ram = "%4.2fG" % (self.cache_size() / GIG)
+        vram = "%4.2fG" % (torch.cuda.memory_allocated() / GB)
+        ram = "%4.2fG" % (self.cache_size() / GB)
 
         in_ram_models = 0
         in_vram_models = 0
@@ -353,17 +362,20 @@ class ModelCache(ModelCacheBase[AnyModel]):
                 )
 
     def make_room(self, size: int) -> None:
-        """Make enough room in the cache to accommodate a new model of indicated size."""
-        # calculate how much memory this model will require
-        # multiplier = 2 if self.precision==torch.float32 else 1
+        """Make enough room in the cache to accommodate a new model of indicated size.
+
+        Note: This function deletes all of the cache's internal references to a model in order to free it. If there are
+        external references to the model, there's nothing that the cache can do about it, and those models will not be
+        garbage-collected.
+        """
         bytes_needed = size
-        maximum_size = self.max_cache_size * GIG  # stored in GB, convert to bytes
+        maximum_size = self.max_cache_size * GB  # stored in GB, convert to bytes
         current_size = self.cache_size()
 
         if current_size + bytes_needed > maximum_size:
             self.logger.debug(
-                f"Max cache size exceeded: {(current_size/GIG):.2f}/{self.max_cache_size:.2f} GB, need an additional"
-                f" {(bytes_needed/GIG):.2f} GB"
+                f"Max cache size exceeded: {(current_size/GB):.2f}/{self.max_cache_size:.2f} GB, need an additional"
+                f" {(bytes_needed/GB):.2f} GB"
             )
 
         self.logger.debug(f"Before making_room: cached_models={len(self._cached_models)}")
@@ -380,7 +392,7 @@ class ModelCache(ModelCacheBase[AnyModel]):
 
             if not cache_entry.locked:
                 self.logger.debug(
-                    f"Removing {model_key} from RAM cache to free at least {(size/GIG):.2f} GB (-{(cache_entry.size/GIG):.2f} GB)"
+                    f"Removing {model_key} from RAM cache to free at least {(size/GB):.2f} GB (-{(cache_entry.size/GB):.2f} GB)"
                 )
                 current_size -= cache_entry.size
                 models_cleared += 1

invokeai/backend/model_manager/load/model_loaders/flux.py

@@ -32,6 +32,7 @@ from invokeai.backend.model_manager.config import (
 )
 from invokeai.backend.model_manager.load.load_default import ModelLoader
 from invokeai.backend.model_manager.load.model_loader_registry import ModelLoaderRegistry
+from invokeai.backend.model_manager.util.model_util import convert_bundle_to_flux_transformer_checkpoint
 from invokeai.backend.util.silence_warnings import SilenceWarnings
 
 try:
@@ -190,6 +191,8 @@ class FluxCheckpointModel(ModelLoader):
         with SilenceWarnings():
             model = Flux(params[config.config_path])
             sd = load_file(model_path)
+            if "model.diffusion_model.double_blocks.0.img_attn.norm.key_norm.scale" in sd:
+                sd = convert_bundle_to_flux_transformer_checkpoint(sd)
             model.load_state_dict(sd, assign=True)
         return model
 
@@ -230,5 +233,7 @@ class FluxBnbQuantizednf4bCheckpointModel(ModelLoader):
                 model = Flux(params[config.config_path])
                 model = quantize_model_nf4(model, modules_to_not_convert=set(), compute_dtype=torch.bfloat16)
             sd = load_file(model_path)
+            if "model.diffusion_model.double_blocks.0.img_attn.norm.key_norm.scale" in sd:
+                sd = convert_bundle_to_flux_transformer_checkpoint(sd)
             model.load_state_dict(sd, assign=True)
         return model

invokeai/backend/model_manager/probe.py

@@ -108,6 +108,8 @@ class ModelProbe(object):
         "CLIPVisionModelWithProjection": ModelType.CLIPVision,
         "T2IAdapter": ModelType.T2IAdapter,
         "CLIPModel": ModelType.CLIPEmbed,
+        "CLIPTextModel": ModelType.CLIPEmbed,
+        "T5EncoderModel": ModelType.T5Encoder,
     }
 
     @classmethod
@@ -224,7 +226,18 @@ class ModelProbe(object):
         ckpt = ckpt.get("state_dict", ckpt)
 
         for key in [str(k) for k in ckpt.keys()]:
-            if key.startswith(("cond_stage_model.", "first_stage_model.", "model.diffusion_model.", "double_blocks.")):
+            if key.startswith(
+                (
+                    "cond_stage_model.",
+                    "first_stage_model.",
+                    "model.diffusion_model.",
+                    # FLUX models in the official BFL format contain keys with the "double_blocks." prefix.
+                    "double_blocks.",
+                    # Some FLUX checkpoint files contain transformer keys prefixed with "model.diffusion_model".
+                    # This prefix is typically used to distinguish between multiple models bundled in a single file.
+                    "model.diffusion_model.double_blocks.",
+                )
+            ):
                 # Keys starting with double_blocks are associated with Flux models
                 return ModelType.Main
             elif key.startswith(("encoder.conv_in", "decoder.conv_in")):
@@ -283,9 +296,16 @@ class ModelProbe(object):
         if (folder_path / "image_encoder.txt").exists():
             return ModelType.IPAdapter
 
-        i = folder_path / "model_index.json"
-        c = folder_path / "config.json"
-        config_path = i if i.exists() else c if c.exists() else None
+        config_path = None
+        for p in [
+            folder_path / "model_index.json",  # pipeline
+            folder_path / "config.json",  # most diffusers
+            folder_path / "text_encoder_2" / "config.json",  # T5 text encoder
+            folder_path / "text_encoder" / "config.json",  # T5 CLIP
+        ]:
+            if p.exists():
+                config_path = p
+                break
 
         if config_path:
             with open(config_path, "r") as file:
@@ -328,7 +348,10 @@ class ModelProbe(object):
                 # TODO: Decide between dev/schnell
                 checkpoint = ModelProbe._scan_and_load_checkpoint(model_path)
                 state_dict = checkpoint.get("state_dict") or checkpoint
-                if "guidance_in.out_layer.weight" in state_dict:
+                if (
+                    "guidance_in.out_layer.weight" in state_dict
+                    or "model.diffusion_model.guidance_in.out_layer.weight" in state_dict
+                ):
                     # For flux, this is a key in invokeai.backend.flux.util.params
                     #   Due to model type and format being the descriminator for model configs this
                     #   is used rather than attempting to support flux with separate model types and format
@@ -336,7 +359,7 @@ class ModelProbe(object):
                     config_file = "flux-dev"
                 else:
                     # For flux, this is a key in invokeai.backend.flux.util.params
-                    #   Due to model type and format being the descriminator for model configs this
+                    #   Due to model type and format being the discriminator for model configs this
                     #   is used rather than attempting to support flux with separate model types and format
                     #   If changed in the future, please fix me
                     config_file = "flux-schnell"
@@ -443,7 +466,10 @@ class CheckpointProbeBase(ProbeBase):
 
     def get_format(self) -> ModelFormat:
         state_dict = self.checkpoint.get("state_dict") or self.checkpoint
-        if "double_blocks.0.img_attn.proj.weight.quant_state.bitsandbytes__nf4" in state_dict:
+        if (
+            "double_blocks.0.img_attn.proj.weight.quant_state.bitsandbytes__nf4" in state_dict
+            or "model.diffusion_model.double_blocks.0.img_attn.proj.weight.quant_state.bitsandbytes__nf4" in state_dict
+        ):
             return ModelFormat.BnbQuantizednf4b
         return ModelFormat("checkpoint")
 
@@ -470,7 +496,10 @@ class PipelineCheckpointProbe(CheckpointProbeBase):
     def get_base_type(self) -> BaseModelType:
         checkpoint = self.checkpoint
         state_dict = self.checkpoint.get("state_dict") or checkpoint
-        if "double_blocks.0.img_attn.norm.key_norm.scale" in state_dict:
+        if (
+            "double_blocks.0.img_attn.norm.key_norm.scale" in state_dict
+            or "model.diffusion_model.double_blocks.0.img_attn.norm.key_norm.scale" in state_dict
+        ):
             return BaseModelType.Flux
         key_name = "model.diffusion_model.input_blocks.2.1.transformer_blocks.0.attn2.to_k.weight"
         if key_name in state_dict and state_dict[key_name].shape[-1] == 768:
@@ -747,8 +776,27 @@ class TextualInversionFolderProbe(FolderProbeBase):
 
 
 class T5EncoderFolderProbe(FolderProbeBase):
+    def get_base_type(self) -> BaseModelType:
+        return BaseModelType.Any
+
     def get_format(self) -> ModelFormat:
-        return ModelFormat.T5Encoder
+        path = self.model_path / "text_encoder_2"
+        if (path / "model.safetensors.index.json").exists():
+            return ModelFormat.T5Encoder
+        files = list(path.glob("*.safetensors"))
+        if len(files) == 0:
+            raise InvalidModelConfigException(f"{self.model_path.as_posix()}: no .safetensors files found")
+
+        # shortcut: look for the quantization in the name
+        if any(x for x in files if "llm_int8" in x.as_posix()):
+            return ModelFormat.BnbQuantizedLlmInt8b
+
+        # more reliable path: probe contents for a 'SCB' key
+        ckpt = read_checkpoint_meta(files[0], scan=True)
+        if any("SCB" in x for x in ckpt.keys()):
+            return ModelFormat.BnbQuantizedLlmInt8b
+
+        raise InvalidModelConfigException(f"{self.model_path.as_posix()}: unknown model format")
 
 
 class ONNXFolderProbe(PipelineFolderProbe):

invokeai/backend/model_manager/util/model_util.py

@@ -133,3 +133,29 @@ def lora_token_vector_length(checkpoint: Dict[str, torch.Tensor]) -> Optional[in
             break
 
     return lora_token_vector_length
+
+
+def convert_bundle_to_flux_transformer_checkpoint(
+    transformer_state_dict: dict[str, torch.Tensor],
+) -> dict[str, torch.Tensor]:
+    original_state_dict: dict[str, torch.Tensor] = {}
+    keys_to_remove: list[str] = []
+
+    for k, v in transformer_state_dict.items():
+        if not k.startswith("model.diffusion_model"):
+            keys_to_remove.append(k)  # This can be removed in the future if we only want to delete transformer keys
+            continue
+        if k.endswith("scale"):
+            # Scale math must be done at bfloat16 due to our current flux model
+            # support limitations at inference time
+            v = v.to(dtype=torch.bfloat16)
+        new_key = k.replace("model.diffusion_model.", "")
+        original_state_dict[new_key] = v
+        keys_to_remove.append(k)
+
+    # Remove processed keys from the original dictionary, leaving others in case
+    # other model state dicts need to be pulled
+    for k in keys_to_remove:
+        del transformer_state_dict[k]
+
+    return original_state_dict

invokeai/backend/quantization/bnb_llm_int8.py

@@ -54,8 +54,10 @@ class InvokeLinear8bitLt(bnb.nn.Linear8bitLt):
 
         # See `bnb.nn.Linear8bitLt._save_to_state_dict()` for the serialization logic of SCB and weight_format.
         scb = state_dict.pop(prefix + "SCB", None)
-        # weight_format is unused, but we pop it so we can validate that there are no unexpected keys.
-        _weight_format = state_dict.pop(prefix + "weight_format", None)
+
+        # Currently, we only support weight_format=0.
+        weight_format = state_dict.pop(prefix + "weight_format", None)
+        assert weight_format == 0
 
         # TODO(ryand): Technically, we should be using `strict`, `missing_keys`, `unexpected_keys`, and `error_msgs`
         # rather than raising an exception to correctly implement this API.
@@ -89,6 +91,14 @@ class InvokeLinear8bitLt(bnb.nn.Linear8bitLt):
             )
             self.bias = bias if bias is None else torch.nn.Parameter(bias)
 
+        # Reset the state. The persisted fields are based on the initialization behaviour in
+        # `bnb.nn.Linear8bitLt.__init__()`.
+        new_state = bnb.MatmulLtState()
+        new_state.threshold = self.state.threshold
+        new_state.has_fp16_weights = False
+        new_state.use_pool = self.state.use_pool
+        self.state = new_state
+
 
 def _convert_linear_layers_to_llm_8bit(
     module: torch.nn.Module, ignore_modules: set[str], outlier_threshold: float, prefix: str = ""

invokeai/backend/stable_diffusion/diffusion/conditioning_data.py

@@ -43,6 +43,11 @@ class FLUXConditioningInfo:
     clip_embeds: torch.Tensor
     t5_embeds: torch.Tensor
 
+    def to(self, device: torch.device | None = None, dtype: torch.dtype | None = None):
+        self.clip_embeds = self.clip_embeds.to(device=device, dtype=dtype)
+        self.t5_embeds = self.t5_embeds.to(device=device, dtype=dtype)
+        return self
+
 
 @dataclass
 class ConditioningFieldData:

invokeai/backend/util/__init__.py

@@ -3,10 +3,9 @@ Initialization file for invokeai.backend.util
 """
 
 from invokeai.backend.util.logging import InvokeAILogger
-from invokeai.backend.util.util import GIG, Chdir, directory_size
+from invokeai.backend.util.util import Chdir, directory_size
 
 __all__ = [
-    "GIG",
     "directory_size",
     "Chdir",
     "InvokeAILogger",

invokeai/backend/util/util.py

@@ -7,9 +7,6 @@ from pathlib import Path
 
 from PIL import Image
 
-# actual size of a gig
-GIG = 1073741824
-
 
 def slugify(value: str, allow_unicode: bool = False) -> str:
     """

invokeai/frontend/web/package.json

@@ -136,6 +136,7 @@
     "@vitest/coverage-v8": "^1.5.0",
     "@vitest/ui": "^1.5.0",
     "concurrently": "^8.2.2",
+    "csstype": "^3.1.3",
     "dpdm": "^3.14.0",
     "eslint": "^8.57.0",
     "eslint-plugin-i18next": "^6.0.9",

invokeai/frontend/web/pnpm-lock.yaml

@@ -238,6 +238,9 @@ devDependencies:
   concurrently:
     specifier: ^8.2.2
     version: 8.2.2
+  csstype:
+    specifier: ^3.1.3
+    version: 3.1.3
   dpdm:
     specifier: ^3.14.0
     version: 3.14.0

invokeai/frontend/web/public/locales/en.json

@@ -164,10 +164,10 @@
         "alpha": "Alpha",
         "selected": "Selected",
         "tab": "Tab",
-        "viewing": "Viewing",
-        "viewingDesc": "Review images in a large gallery view",
-        "editing": "Editing",
-        "editingDesc": "Edit on the Control Layers canvas",
+        "view": "View",
+        "viewDesc": "Review images in a large gallery view",
+        "edit": "Edit",
+        "editDesc": "Edit on the Canvas",
         "comparing": "Comparing",
         "comparingDesc": "Comparing two images",
         "enabled": "Enabled",
@@ -328,9 +328,13 @@
         "completedIn": "Completed in",
         "batch": "Batch",
         "origin": "Origin",
-        "originCanvas": "Canvas",
-        "originWorkflows": "Workflows",
-        "originOther": "Other",
+        "destination": "Destination",
+        "upscaling": "Upscaling",
+        "canvas": "Canvas",
+        "generation": "Generation",
+        "workflows": "Workflows",
+        "other": "Other",
+        "gallery": "Gallery",
         "batchFieldValues": "Batch Field Values",
         "item": "Item",
         "session": "Session",
@@ -702,6 +706,8 @@
         "availableModels": "Available Models",
         "baseModel": "Base Model",
         "cancel": "Cancel",
+        "clipEmbed": "CLIP Embed",
+        "clipVision": "CLIP Vision",
         "config": "Config",
         "convert": "Convert",
         "convertingModelBegin": "Converting Model. Please wait.",
@@ -789,6 +795,7 @@
         "settings": "Settings",
         "simpleModelPlaceholder": "URL or path to a local file or diffusers folder",
         "source": "Source",
+        "spandrelImageToImage": "Image to Image (Spandrel)",
         "starterModels": "Starter Models",
         "starterModelsInModelManager": "Starter Models can be found in Model Manager",
         "syncModels": "Sync Models",
@@ -797,6 +804,7 @@
         "loraTriggerPhrases": "LoRA Trigger Phrases",
         "mainModelTriggerPhrases": "Main Model Trigger Phrases",
         "typePhraseHere": "Type phrase here",
+        "t5Encoder": "T5 Encoder",
         "upcastAttention": "Upcast Attention",
         "uploadImage": "Upload Image",
         "urlOrLocalPath": "URL or Local Path",
@@ -1646,6 +1654,15 @@
         "storeNotInitialized": "Store is not initialized"
     },
     "controlLayers": {
+        "bookmark": "Bookmark for Quick Switch",
+        "removeBookmark": "Remove Bookmark",
+        "saveCanvasToGallery": "Save Canvas To Gallery",
+        "saveBboxToGallery": "Save Bbox To Gallery",
+        "savedToGalleryOk": "Saved to Gallery",
+        "savedToGalleryError": "Error saving to gallery",
+        "mergeVisible": "Merge Visible",
+        "mergeVisibleOk": "Merged visible layers",
+        "mergeVisibleError": "Error merging visible layers",
         "clearHistory": "Clear History",
         "generateMode": "Generate",
         "generateModeDesc": "Create individual images. Generated images are added directly to the gallery.",
@@ -1657,6 +1674,7 @@
         "clearCaches": "Clear Caches",
         "recalculateRects": "Recalculate Rects",
         "clipToBbox": "Clip Strokes to Bbox",
+        "compositeMaskedRegions": "Composite Masked Regions",
         "addLayer": "Add Layer",
         "duplicate": "Duplicate",
         "moveToFront": "Move to Front",
@@ -1675,35 +1693,47 @@
         "deletePrompt": "Delete Prompt",
         "resetRegion": "Reset Region",
         "debugLayers": "Debug Layers",
+        "showHUD": "Show HUD",
         "rectangle": "Rectangle",
         "maskFill": "Mask Fill",
         "addPositivePrompt": "Add $t(common.positivePrompt)",
         "addNegativePrompt": "Add $t(common.negativePrompt)",
         "addIPAdapter": "Add $t(common.ipAdapter)",
+        "addRasterLayer": "Add $t(controlLayers.rasterLayer)",
+        "addControlLayer": "Add $t(controlLayers.controlLayer)",
+        "addInpaintMask": "Add $t(controlLayers.inpaintMask)",
+        "addRegionalGuidance": "Add $t(controlLayers.regionalGuidance)",
         "regionalGuidanceLayer": "$t(controlLayers.regionalGuidance) $t(unifiedCanvas.layer)",
         "raster": "Raster",
-        "rasterLayer_one": "Raster Layer",
-        "controlLayer_one": "Control Layer",
-        "inpaintMask_one": "Inpaint Mask",
-        "regionalGuidance_one": "Regional Guidance",
-        "ipAdapter_one": "IP Adapter",
-        "rasterLayer_other": "Raster Layers",
-        "controlLayer_other": "Control Layers",
-        "inpaintMask_other": "Inpaint Masks",
-        "regionalGuidance_other": "Regional Guidance",
-        "ipAdapter_other": "IP Adapters",
+        "rasterLayer": "Raster Layer",
+        "controlLayer": "Control Layer",
+        "inpaintMask": "Inpaint Mask",
+        "regionalGuidance": "Regional Guidance",
+        "ipAdapter": "IP Adapter",
+        "sendToGallery": "Send To Gallery",
+        "sendToGalleryDesc": "Generations will be sent to the gallery.",
+        "sendToCanvas": "Send To Canvas",
+        "sendToCanvasDesc": "Generations will be staged onto the canvas.",
+        "rasterLayer_withCount_one": "$t(controlLayers.rasterLayer)",
+        "controlLayer_withCount_one": "$t(controlLayers.controlLayer)",
+        "inpaintMask_withCount_one": "$t(controlLayers.inpaintMask)",
+        "regionalGuidance_withCount_one": "$t(controlLayers.regionalGuidance)",
+        "ipAdapter_withCount_one": "$t(controlLayers.ipAdapter)",
+        "rasterLayer_withCount_other": "Raster Layers",
+        "controlLayer_withCount_other": "Control Layers",
+        "inpaintMask_withCount_other": "Inpaint Masks",
+        "regionalGuidance_withCount_other": "Regional Guidance",
+        "ipAdapter_withCount_other": "IP Adapters",
         "opacity": "Opacity",
         "regionalGuidance_withCount_hidden": "Regional Guidance ({{count}} hidden)",
-        "controlAdapters_withCount_hidden": "Control Adapters ({{count}} hidden)",
         "controlLayers_withCount_hidden": "Control Layers ({{count}} hidden)",
         "rasterLayers_withCount_hidden": "Raster Layers ({{count}} hidden)",
-        "ipAdapters_withCount_hidden": "IP Adapters ({{count}} hidden)",
+        "globalIPAdapters_withCount_hidden": "Global IP Adapters ({{count}} hidden)",
         "inpaintMasks_withCount_hidden": "Inpaint Masks ({{count}} hidden)",
         "regionalGuidance_withCount_visible": "Regional Guidance ({{count}})",
-        "controlAdapters_withCount_visible": "Control Adapters ({{count}})",
         "controlLayers_withCount_visible": "Control Layers ({{count}})",
         "rasterLayers_withCount_visible": "Raster Layers ({{count}})",
-        "ipAdapters_withCount_visible": "IP Adapters ({{count}})",
+        "globalIPAdapters_withCount_visible": "Global IP Adapters ({{count}})",
         "inpaintMasks_withCount_visible": "Inpaint Masks ({{count}})",
         "globalControlAdapter": "Global $t(controlnet.controlAdapter_one)",
         "globalControlAdapterLayer": "Global $t(controlnet.controlAdapter_one) $t(unifiedCanvas.layer)",
@@ -1716,8 +1746,8 @@
         "clearProcessor": "Clear Processor",
         "resetProcessor": "Reset Processor to Defaults",
         "noLayersAdded": "No Layers Added",
-        "layers_one": "Layer",
-        "layers_other": "Layers",
+        "layer_one": "Layer",
+        "layer_other": "Layers",
         "objects_zero": "empty",
         "objects_one": "{{count}} object",
         "objects_other": "{{count}} objects",
@@ -1737,6 +1767,7 @@
         "flipHorizontal": "Flip Horizontal",
         "flipVertical": "Flip Vertical",
         "fill": {
+            "fillColor": "Fill Color",
             "fillStyle": "Fill Style",
             "solid": "Solid",
             "grid": "Grid",
@@ -1752,7 +1783,6 @@
             "bbox": "Bbox",
             "move": "Move",
             "view": "View",
-            "transform": "Transform",
             "colorPicker": "Color Picker"
         },
         "filter": {
@@ -1762,6 +1792,13 @@
             "preview": "Preview",
             "apply": "Apply",
             "cancel": "Cancel"
+        },
+        "transform": {
+            "transform": "Transform",
+            "fitToBbox": "Fit to Bbox",
+            "reset": "Reset",
+            "apply": "Apply",
+            "cancel": "Cancel"
         }
     },
     "upscaling": {

invokeai/frontend/web/src/app/components/App.tsx

@@ -16,6 +16,7 @@ import { DynamicPromptsModal } from 'features/dynamicPrompts/components/DynamicP
 import { useStarterModelsToast } from 'features/modelManagerV2/hooks/useStarterModelsToast';
 import { ClearQueueConfirmationsAlertDialog } from 'features/queue/components/ClearQueueConfirmationAlertDialog';
 import { StylePresetModal } from 'features/stylePresets/components/StylePresetForm/StylePresetModal';
+import { activeStylePresetIdChanged } from 'features/stylePresets/store/stylePresetSlice';
 import RefreshAfterResetModal from 'features/system/components/SettingsModal/RefreshAfterResetModal';
 import SettingsModal from 'features/system/components/SettingsModal/SettingsModal';
 import { configChanged } from 'features/system/store/configSlice';
@@ -43,10 +44,17 @@ interface Props {
     action: 'sendToImg2Img' | 'sendToCanvas' | 'useAllParameters';
   };
   selectedWorkflowId?: string;
-  destination?: TabName | undefined;
+  selectedStylePresetId?: string;
+  destination?: TabName;
 }
 
-const App = ({ config = DEFAULT_CONFIG, selectedImage, selectedWorkflowId, destination }: Props) => {
+const App = ({
+  config = DEFAULT_CONFIG,
+  selectedImage,
+  selectedWorkflowId,
+  selectedStylePresetId,
+  destination,
+}: Props) => {
   const language = useAppSelector(selectLanguage);
   const logger = useLogger('system');
   const dispatch = useAppDispatch();
@@ -85,6 +93,12 @@ const App = ({ config = DEFAULT_CONFIG, selectedImage, selectedWorkflowId, desti
     }
   }, [selectedWorkflowId, getAndLoadWorkflow]);
 
+  useEffect(() => {
+    if (selectedStylePresetId) {
+      dispatch(activeStylePresetIdChanged(selectedStylePresetId));
+    }
+  }, [dispatch, selectedStylePresetId]);
+
   useEffect(() => {
     if (destination) {
       dispatch(setActiveTab(destination));

invokeai/frontend/web/src/app/components/InvokeAIUI.tsx

@@ -45,6 +45,7 @@ interface Props extends PropsWithChildren {
     action: 'sendToImg2Img' | 'sendToCanvas' | 'useAllParameters';
   };
   selectedWorkflowId?: string;
+  selectedStylePresetId?: string;
   destination?: TabName;
   customStarUi?: CustomStarUi;
   socketOptions?: Partial<ManagerOptions & SocketOptions>;
@@ -66,6 +67,7 @@ const InvokeAIUI = ({
   queueId,
   selectedImage,
   selectedWorkflowId,
+  selectedStylePresetId,
   destination,
   customStarUi,
   socketOptions,
@@ -227,6 +229,7 @@ const InvokeAIUI = ({
                 config={config}
                 selectedImage={selectedImage}
                 selectedWorkflowId={selectedWorkflowId}
+                selectedStylePresetId={selectedStylePresetId}
                 destination={destination}
               />
             </AppDndContext>

invokeai/frontend/web/src/app/store/middleware/listenerMiddleware/listeners/addCommitStagingAreaImageListener.ts

@@ -68,7 +68,7 @@ export const addStagingListeners = (startAppListening: AppStartListening) => {
         objects: [imageObject],
       };
 
-      api.dispatch(rasterLayerAdded({ overrides, isSelected: true }));
+      api.dispatch(rasterLayerAdded({ overrides, isSelected: false }));
       api.dispatch(sessionStagingAreaReset());
     },
   });

invokeai/frontend/web/src/app/store/middleware/listenerMiddleware/listeners/enqueueRequestedLinear.ts

@@ -31,7 +31,7 @@ export const addEnqueueRequestedLinear = (startAppListening: AppStartListening)
 
       let didStartStaging = false;
 
-      if (!state.canvasSession.isStaging && state.canvasSession.mode === 'compose') {
+      if (!state.canvasSession.isStaging && state.canvasSettings.sendToCanvas) {
         dispatch(sessionStartedStaging());
         didStartStaging = true;
       }
@@ -70,7 +70,11 @@ export const addEnqueueRequestedLinear = (startAppListening: AppStartListening)
 
       const { g, noise, posCond } = buildGraphResult.value;
 
-      const prepareBatchResult = withResult(() => prepareLinearUIBatch(state, g, prepend, noise, posCond));
+      const destination = state.canvasSettings.sendToCanvas ? 'canvas' : 'gallery';
+
+      const prepareBatchResult = withResult(() =>
+        prepareLinearUIBatch(state, g, prepend, noise, posCond, 'generation', destination)
+      );
 
       if (isErr(prepareBatchResult)) {
         log.error({ error: serializeError(prepareBatchResult.error) }, 'Failed to prepare batch');

invokeai/frontend/web/src/app/store/middleware/listenerMiddleware/listeners/enqueueRequestedNodes.ts

@@ -32,6 +32,7 @@ export const addEnqueueRequestedNodes = (startAppListening: AppStartListening) =
           workflow: builtWorkflow,
           runs: state.params.iterations,
           origin: 'workflows',
+          destination: 'gallery',
         },
         prepend: action.payload.prepend,
       };

invokeai/frontend/web/src/app/store/middleware/listenerMiddleware/listeners/enqueueRequestedUpscale.ts

@@ -16,7 +16,7 @@ export const addEnqueueRequestedUpscale = (startAppListening: AppStartListening)
 
       const { g, noise, posCond } = await buildMultidiffusionUpscaleGraph(state);
 
-      const batchConfig = prepareLinearUIBatch(state, g, prepend, noise, posCond);
+      const batchConfig = prepareLinearUIBatch(state, g, prepend, noise, posCond, 'upscaling', 'gallery');
 
       const req = dispatch(
         queueApi.endpoints.enqueueBatch.initiate(batchConfig, {

invokeai/frontend/web/src/app/store/store.ts

@@ -11,7 +11,6 @@ import { canvasSettingsPersistConfig, canvasSettingsSlice } from 'features/contr
 import { canvasPersistConfig, canvasSlice, canvasUndoableConfig } from 'features/controlLayers/store/canvasSlice';
 import { lorasPersistConfig, lorasSlice } from 'features/controlLayers/store/lorasSlice';
 import { paramsPersistConfig, paramsSlice } from 'features/controlLayers/store/paramsSlice';
-import { toolPersistConfig, toolSlice } from 'features/controlLayers/store/toolSlice';
 import { deleteImageModalSlice } from 'features/deleteImageModal/store/slice';
 import { dynamicPromptsPersistConfig, dynamicPromptsSlice } from 'features/dynamicPrompts/store/dynamicPromptsSlice';
 import { galleryPersistConfig, gallerySlice } from 'features/gallery/store/gallerySlice';
@@ -63,7 +62,6 @@ const allReducers = {
   [upscaleSlice.name]: upscaleSlice.reducer,
   [stylePresetSlice.name]: stylePresetSlice.reducer,
   [paramsSlice.name]: paramsSlice.reducer,
-  [toolSlice.name]: toolSlice.reducer,
   [canvasSettingsSlice.name]: canvasSettingsSlice.reducer,
   [canvasSessionSlice.name]: canvasSessionSlice.reducer,
   [lorasSlice.name]: lorasSlice.reducer,
@@ -109,7 +107,6 @@ const persistConfigs: { [key in keyof typeof allReducers]?: PersistConfig } = {
   [upscalePersistConfig.name]: upscalePersistConfig,
   [stylePresetPersistConfig.name]: stylePresetPersistConfig,
   [paramsPersistConfig.name]: paramsPersistConfig,
-  [toolPersistConfig.name]: toolPersistConfig,
   [canvasSettingsPersistConfig.name]: canvasSettingsPersistConfig,
   [canvasSessionPersistConfig.name]: canvasSessionPersistConfig,
   [lorasPersistConfig.name]: lorasPersistConfig,

invokeai/frontend/web/src/common/components/IconSwitch.tsx

@@ -0,0 +1,104 @@
+import type { SystemStyleObject } from '@invoke-ai/ui-library';
+import { Box, Flex, IconButton, Tooltip, useToken } from '@invoke-ai/ui-library';
+import type { ReactElement, ReactNode } from 'react';
+import { memo, useCallback, useMemo } from 'react';
+
+type IconSwitchProps = {
+  isChecked: boolean;
+  onChange: (checked: boolean) => void;
+  iconChecked: ReactElement;
+  tooltipChecked?: ReactNode;
+  iconUnchecked: ReactElement;
+  tooltipUnchecked?: ReactNode;
+  ariaLabel: string;
+};
+
+const getSx = (padding: string | number): SystemStyleObject => ({
+  transition: 'left 0.1s ease-in-out, transform 0.1s ease-in-out',
+  '&[data-checked="true"]': {
+    left: `calc(100% - ${padding})`,
+    transform: 'translateX(-100%)',
+  },
+  '&[data-checked="false"]': {
+    left: padding,
+    transform: 'translateX(0)',
+  },
+});
+
+export const IconSwitch = memo(
+  ({
+    isChecked,
+    onChange,
+    iconChecked,
+    tooltipChecked,
+    iconUnchecked,
+    tooltipUnchecked,
+    ariaLabel,
+  }: IconSwitchProps) => {
+    const onUncheck = useCallback(() => {
+      onChange(false);
+    }, [onChange]);
+    const onCheck = useCallback(() => {
+      onChange(true);
+    }, [onChange]);
+
+    const gap = useToken('space', 1.5);
+    const sx = useMemo(() => getSx(gap), [gap]);
+
+    return (
+      <Flex
+        position="relative"
+        bg="base.800"
+        borderRadius="base"
+        alignItems="center"
+        justifyContent="center"
+        h="full"
+        p={gap}
+        gap={gap}
+      >
+        <Box
+          position="absolute"
+          borderRadius="base"
+          bg="invokeBlue.400"
+          w={12}
+          top={gap}
+          bottom={gap}
+          data-checked={isChecked}
+          sx={sx}
+        />
+        <Tooltip hasArrow label={tooltipUnchecked}>
+          <IconButton
+            size="sm"
+            fontSize={16}
+            icon={iconUnchecked}
+            onClick={onUncheck}
+            variant={!isChecked ? 'solid' : 'ghost'}
+            colorScheme={!isChecked ? 'invokeBlue' : 'base'}
+            aria-label={ariaLabel}
+            data-checked={!isChecked}
+            w={12}
+            alignSelf="stretch"
+            h="auto"
+          />
+        </Tooltip>
+        <Tooltip hasArrow label={tooltipChecked}>
+          <IconButton
+            size="sm"
+            fontSize={16}
+            icon={iconChecked}
+            onClick={onCheck}
+            variant={isChecked ? 'solid' : 'ghost'}
+            colorScheme={isChecked ? 'invokeBlue' : 'base'}
+            aria-label={ariaLabel}
+            data-checked={isChecked}
+            w={12}
+            alignSelf="stretch"
+            h="auto"
+          />
+        </Tooltip>
+      </Flex>
+    );
+  }
+);
+
+IconSwitch.displayName = 'IconSwitch';

invokeai/frontend/web/src/common/hooks/useBoolean.ts

@@ -1,52 +1,74 @@
+import { useStore } from '@nanostores/react';
 import type { WritableAtom } from 'nanostores';
-import { useCallback, useMemo, useState } from 'react';
+import { atom } from 'nanostores';
+import { useCallback, useState } from 'react';
 
-export const useBoolean = (initialValue: boolean) => {
-  const [isTrue, set] = useState(initialValue);
-  const setTrue = useCallback(() => set(true), []);
-  const setFalse = useCallback(() => set(false), []);
-  const toggle = useCallback(() => set((v) => !v), []);
+type UseBoolean = {
+  isTrue: boolean;
+  setTrue: () => void;
+  setFalse: () => void;
+  set: (value: boolean) => void;
+  toggle: () => void;
+};
 
-  const api = useMemo(
-    () => ({
+/**
+ * Creates a hook to manage a boolean state. The boolean is stored in a nanostores atom.
+ * Returns a tuple containing the hook and the atom. Use this for global boolean state.
+ * @param initialValue Initial value of the boolean
+ */
+export const buildUseBoolean = (initialValue: boolean): [() => UseBoolean, WritableAtom<boolean>] => {
+  const $boolean = atom(initialValue);
+
+  const setTrue = () => {
+    $boolean.set(true);
+  };
+  const setFalse = () => {
+    $boolean.set(false);
+  };
+  const set = (value: boolean) => {
+    $boolean.set(value);
+  };
+  const toggle = () => {
+    $boolean.set(!$boolean.get());
+  };
+
+  const useBoolean = () => {
+    const isTrue = useStore($boolean);
+
+    return {
       isTrue,
-      set,
       setTrue,
       setFalse,
+      set,
       toggle,
-    }),
-    [isTrue, set, setTrue, setFalse, toggle]
-  );
+    };
+  };
 
-  return api;
+  return [useBoolean, $boolean] as const;
 };
 
-export const buildUseBoolean = ($boolean: WritableAtom<boolean>) => {
-  return () => {
-    const setTrue = useCallback(() => {
-      $boolean.set(true);
-    }, []);
-    const setFalse = useCallback(() => {
-      $boolean.set(false);
-    }, []);
-    const set = useCallback((value: boolean) => {
-      $boolean.set(value);
-    }, []);
-    const toggle = useCallback(() => {
-      $boolean.set(!$boolean.get());
-    }, []);
+/**
+ * Hook to manage a boolean state. Use this for a local boolean state.
+ * @param initialValue Initial value of the boolean
+ */
+export const useBoolean = (initialValue: boolean) => {
+  const [isTrue, set] = useState(initialValue);
 
-    const api = useMemo(
-      () => ({
-        setTrue,
-        setFalse,
-        set,
-        toggle,
-        $boolean,
-      }),
-      [set, setFalse, setTrue, toggle]
-    );
+  const setTrue = useCallback(() => {
+    set(true);
+  }, [set]);
+  const setFalse = useCallback(() => {
+    set(false);
+  }, [set]);
+  const toggle = useCallback(() => {
+    set((val) => !val);
+  }, [set]);
 
-    return api;
+  return {
+    isTrue,
+    setTrue,
+    setFalse,
+    set,
+    toggle,
   };
 };

invokeai/frontend/web/src/common/hooks/useGlobalHotkeys.ts

@@ -1,7 +1,7 @@
 import { useAppDispatch } from 'app/store/storeHooks';
 import { addScope, removeScope, setScopes } from 'common/hooks/interactionScopes';
+import { useClearQueue } from 'features/queue/components/ClearQueueConfirmationAlertDialog';
 import { useCancelCurrentQueueItem } from 'features/queue/hooks/useCancelCurrentQueueItem';
-import { useClearQueue } from 'features/queue/hooks/useClearQueue';
 import { useQueueBack } from 'features/queue/hooks/useQueueBack';
 import { useQueueFront } from 'features/queue/hooks/useQueueFront';
 import { useFeatureStatus } from 'features/system/hooks/useFeatureStatus';

invokeai/frontend/web/src/common/hooks/useIsReadyToEnqueue.ts

@@ -128,7 +128,7 @@ const createSelector = (templates: Templates, isConnected: boolean) =>
         canvas.controlLayers.entities
           .filter((controlLayer) => controlLayer.isEnabled)
           .forEach((controlLayer, i) => {
-            const layerLiteral = i18n.t('controlLayers.layers_one');
+            const layerLiteral = i18n.t('controlLayers.layer_one');
             const layerNumber = i + 1;
             const layerType = i18n.t(LAYER_TYPE_TO_TKEY['control_layer']);
             const prefix = `${layerLiteral} #${layerNumber} (${layerType})`;
@@ -158,7 +158,7 @@ const createSelector = (templates: Templates, isConnected: boolean) =>
         canvas.ipAdapters.entities
           .filter((entity) => entity.isEnabled)
           .forEach((entity, i) => {
-            const layerLiteral = i18n.t('controlLayers.layers_one');
+            const layerLiteral = i18n.t('controlLayers.layer_one');
             const layerNumber = i + 1;
             const layerType = i18n.t(LAYER_TYPE_TO_TKEY[entity.type]);
             const prefix = `${layerLiteral} #${layerNumber} (${layerType})`;
@@ -186,7 +186,7 @@ const createSelector = (templates: Templates, isConnected: boolean) =>
         canvas.regions.entities
           .filter((entity) => entity.isEnabled)
           .forEach((entity, i) => {
-            const layerLiteral = i18n.t('controlLayers.layers_one');
+            const layerLiteral = i18n.t('controlLayers.layer_one');
             const layerNumber = i + 1;
             const layerType = i18n.t(LAYER_TYPE_TO_TKEY[entity.type]);
             const prefix = `${layerLiteral} #${layerNumber} (${layerType})`;
@@ -223,7 +223,7 @@ const createSelector = (templates: Templates, isConnected: boolean) =>
         canvas.rasterLayers.entities
           .filter((entity) => entity.isEnabled)
           .forEach((entity, i) => {
-            const layerLiteral = i18n.t('controlLayers.layers_one');
+            const layerLiteral = i18n.t('controlLayers.layer_one');
             const layerNumber = i + 1;
             const layerType = i18n.t(LAYER_TYPE_TO_TKEY[entity.type]);
             const prefix = `${layerLiteral} #${layerNumber} (${layerType})`;

invokeai/frontend/web/src/features/controlLayers/components/CanvasAddEntityButtons.tsx

@@ -31,22 +31,22 @@ export const CanvasAddEntityButtons = memo(() => {
   }, [dispatch]);
 
   return (
-    <Flex flexDir="column" w="full" h="full" alignItems="center" justifyContent="center">
-      <ButtonGroup orientation="vertical" isAttached={false}>
+    <Flex flexDir="column" w="full" h="full" alignItems="center">
+      <ButtonGroup position="relative" orientation="vertical" isAttached={false} top="20%">
         <Button variant="ghost" justifyContent="flex-start" leftIcon={<PiPlusBold />} onClick={addInpaintMask}>
-          {t('controlLayers.inpaintMask', { count: 1 })}
+          {t('controlLayers.inpaintMask')}
         </Button>
         <Button variant="ghost" justifyContent="flex-start" leftIcon={<PiPlusBold />} onClick={addRegionalGuidance}>
-          {t('controlLayers.regionalGuidance', { count: 1 })}
+          {t('controlLayers.regionalGuidance')}
         </Button>
         <Button variant="ghost" justifyContent="flex-start" leftIcon={<PiPlusBold />} onClick={addRasterLayer}>
-          {t('controlLayers.rasterLayer', { count: 1 })}
+          {t('controlLayers.rasterLayer')}
         </Button>
         <Button variant="ghost" justifyContent="flex-start" leftIcon={<PiPlusBold />} onClick={addControlLayer}>
-          {t('controlLayers.controlLayer', { count: 1 })}
+          {t('controlLayers.controlLayer')}
         </Button>
         <Button variant="ghost" justifyContent="flex-start" leftIcon={<PiPlusBold />} onClick={addIPAdapter}>
-          {t('controlLayers.ipAdapter', { count: 1 })}
+          {t('controlLayers.globalIPAdapter')}
         </Button>
       </ButtonGroup>
     </Flex>

invokeai/frontend/web/src/features/controlLayers/components/CanvasEditor.stories.tsx

@@ -1,6 +1,6 @@
 import { Flex } from '@invoke-ai/ui-library';
 import type { Meta, StoryObj } from '@storybook/react';
-import { CanvasEditor } from 'features/controlLayers/components/ControlLayersEditor';
+import { CanvasEditor } from 'features/controlLayers/components/CanvasEditor';
 
 const meta: Meta<typeof CanvasEditor> = {
   title: 'Feature/ControlLayers',

invokeai/frontend/web/src/features/controlLayers/components/CanvasEditor.tsx

@@ -2,12 +2,12 @@
 import { Flex } from '@invoke-ai/ui-library';
 import { useScopeOnFocus } from 'common/hooks/interactionScopes';
 import { CanvasDropArea } from 'features/controlLayers/components/CanvasDropArea';
-import { ControlLayersToolbar } from 'features/controlLayers/components/ControlLayersToolbar';
 import { Filter } from 'features/controlLayers/components/Filters/Filter';
 import { StageComponent } from 'features/controlLayers/components/StageComponent';
 import { StagingAreaIsStagingGate } from 'features/controlLayers/components/StagingArea/StagingAreaIsStagingGate';
 import { StagingAreaToolbar } from 'features/controlLayers/components/StagingArea/StagingAreaToolbar';
-import { Transform } from 'features/controlLayers/components/Transform';
+import { CanvasToolbar } from 'features/controlLayers/components/Toolbar/CanvasToolbar';
+import { Transform } from 'features/controlLayers/components/Transform/Transform';
 import { CanvasManagerProviderGate } from 'features/controlLayers/contexts/CanvasManagerProviderGate';
 import { memo, useRef } from 'react';
 
@@ -28,16 +28,16 @@ export const CanvasEditor = memo(() => {
       alignItems="center"
       justifyContent="center"
     >
-      <ControlLayersToolbar />
+      <CanvasToolbar />
       <StageComponent />
-      <Flex position="absolute" bottom={8} gap={2} align="center" justify="center">
+      <Flex position="absolute" bottom={4} gap={2} align="center" justify="center">
         <CanvasManagerProviderGate>
           <StagingAreaIsStagingGate>
             <StagingAreaToolbar />
           </StagingAreaIsStagingGate>
         </CanvasManagerProviderGate>
       </Flex>
-      <Flex position="absolute" bottom={8}>
+      <Flex position="absolute" bottom={4}>
         <CanvasManagerProviderGate>
           <Filter />
           <Transform />

invokeai/frontend/web/src/features/controlLayers/components/CanvasEntityList/EntityListActionBarAddLayerMenuItems.tsx

@@ -1,5 +1,6 @@
 import { MenuItem } from '@invoke-ai/ui-library';
 import { useAppDispatch } from 'app/store/storeHooks';
+import { useDefaultIPAdapter } from 'features/controlLayers/hooks/useLayerControlAdapter';
 import {
   controlLayerAdded,
   inpaintMaskAdded,
@@ -14,6 +15,7 @@ import { PiPlusBold } from 'react-icons/pi';
 export const CanvasEntityListMenuItems = memo(() => {
   const { t } = useTranslation();
   const dispatch = useAppDispatch();
+  const defaultIPAdapter = useDefaultIPAdapter();
   const addInpaintMask = useCallback(() => {
     dispatch(inpaintMaskAdded({ isSelected: true }));
   }, [dispatch]);
@@ -27,25 +29,26 @@ export const CanvasEntityListMenuItems = memo(() => {
     dispatch(controlLayerAdded({ isSelected: true }));
   }, [dispatch]);
   const addIPAdapter = useCallback(() => {
-    dispatch(ipaAdded({ isSelected: true }));
-  }, [dispatch]);
+    const overrides = { ipAdapter: defaultIPAdapter };
+    dispatch(ipaAdded({ isSelected: true, overrides }));
+  }, [defaultIPAdapter, dispatch]);
 
   return (
     <>
       <MenuItem icon={<PiPlusBold />} onClick={addInpaintMask}>
-        {t('controlLayers.inpaintMask', { count: 1 })}
+        {t('controlLayers.inpaintMask')}
       </MenuItem>
       <MenuItem icon={<PiPlusBold />} onClick={addRegionalGuidance}>
-        {t('controlLayers.regionalGuidance', { count: 1 })}
+        {t('controlLayers.regionalGuidance')}
       </MenuItem>
       <MenuItem icon={<PiPlusBold />} onClick={addRasterLayer}>
-        {t('controlLayers.rasterLayer', { count: 1 })}
+        {t('controlLayers.rasterLayer')}
       </MenuItem>
       <MenuItem icon={<PiPlusBold />} onClick={addControlLayer}>
-        {t('controlLayers.controlLayer', { count: 1 })}
+        {t('controlLayers.controlLayer')}
       </MenuItem>
       <MenuItem icon={<PiPlusBold />} onClick={addIPAdapter}>
-        {t('controlLayers.ipAdapter', { count: 1 })}
+        {t('controlLayers.globalIPAdapter')}
       </MenuItem>
     </>
   );

invokeai/frontend/web/src/features/controlLayers/components/CanvasEntityList/EntityListActionBarSelectedEntityOpacity.tsx

@@ -22,7 +22,7 @@ import {
   selectEntity,
   selectSelectedEntityIdentifier,
 } from 'features/controlLayers/store/selectors';
-import { isDrawableEntity } from 'features/controlLayers/store/types';
+import { isRenderableEntity } from 'features/controlLayers/store/types';
 import { clamp, round } from 'lodash-es';
 import type { KeyboardEvent } from 'react';
 import { memo, useCallback, useEffect, useState } from 'react';
@@ -37,11 +37,11 @@ function formatPct(v: number | string) {
   return `${round(Number(v), 2).toLocaleString()}%`;
 }
 
-function mapSliderValueToOpacity(value: number) {
+function mapSliderValueToRawValue(value: number) {
   return value / 100;
 }
 
-function mapOpacityToSliderValue(opacity: number) {
+function mapRawValueToSliderValue(opacity: number) {
   return opacity * 100;
 }
 
@@ -50,14 +50,14 @@ function formatSliderValue(value: number) {
 }
 
 const marks = [
-  mapOpacityToSliderValue(0),
-  mapOpacityToSliderValue(0.25),
-  mapOpacityToSliderValue(0.5),
-  mapOpacityToSliderValue(0.75),
-  mapOpacityToSliderValue(1),
+  mapRawValueToSliderValue(0),
+  mapRawValueToSliderValue(0.25),
+  mapRawValueToSliderValue(0.5),
+  mapRawValueToSliderValue(0.75),
+  mapRawValueToSliderValue(1),
 ];
 
-const sliderDefaultValue = mapOpacityToSliderValue(1);
+const sliderDefaultValue = mapRawValueToSliderValue(1);
 
 const snapCandidates = marks.slice(1, marks.length - 1);
 
@@ -70,7 +70,7 @@ const selectOpacity = createSelector(selectCanvasSlice, (canvas) => {
   if (!selectedEntity) {
     return 1; // fallback to 100% opacity
   }
-  if (!isDrawableEntity(selectedEntity)) {
+  if (!isRenderableEntity(selectedEntity)) {
     return 1; // fallback to 100% opacity
   }
   // Opacity is a float from 0-1, but we want to display it as a percentage
@@ -95,7 +95,7 @@ export const SelectedEntityOpacity = memo(() => {
       if (!$shift.get()) {
         snappedOpacity = snapToNearest(opacity, snapCandidates, 2);
       }
-      const mappedOpacity = mapSliderValueToOpacity(snappedOpacity);
+      const mappedOpacity = mapSliderValueToRawValue(snappedOpacity);
 
       dispatch(entityOpacityChanged({ entityIdentifier: selectedEntityIdentifier, opacity: mappedOpacity }));
     },
@@ -157,7 +157,7 @@ export const SelectedEntityOpacity = memo(() => {
             clampValueOnBlur={false}
             variant="outline"
           >
-            <NumberInputField paddingInlineEnd={7} />
+            <NumberInputField paddingInlineEnd={7} _focusVisible={{ zIndex: 0 }} />
             <PopoverTrigger>
               <IconButton
                 aria-label="open-slider"

invokeai/frontend/web/src/features/controlLayers/components/CanvasModeSwitcher.tsx

@@ -1,29 +0,0 @@
-import { Button, ButtonGroup } from '@invoke-ai/ui-library';
-import { createSelector } from '@reduxjs/toolkit';
-import { useAppDispatch, useAppSelector } from 'app/store/storeHooks';
-import { selectCanvasSessionSlice, sessionModeChanged } from 'features/controlLayers/store/canvasSessionSlice';
-import { memo, useCallback } from 'react';
-import { useTranslation } from 'react-i18next';
-
-const selectCanvasMode = createSelector(selectCanvasSessionSlice, (canvasSession) => canvasSession.mode);
-
-export const CanvasModeSwitcher = memo(() => {
-  const { t } = useTranslation();
-  const dispatch = useAppDispatch();
-  const mode = useAppSelector(selectCanvasMode);
-  const onClickGenerate = useCallback(() => dispatch(sessionModeChanged({ mode: 'generate' })), [dispatch]);
-  const onClickCompose = useCallback(() => dispatch(sessionModeChanged({ mode: 'compose' })), [dispatch]);
-
-  return (
-    <ButtonGroup variant="outline">
-      <Button onClick={onClickGenerate} colorScheme={mode === 'generate' ? 'invokeBlue' : 'base'}>
-        {t('controlLayers.generateMode')}
-      </Button>
-      <Button onClick={onClickCompose} colorScheme={mode === 'compose' ? 'invokeBlue' : 'base'}>
-        {t('controlLayers.composeMode')}
-      </Button>
-    </ButtonGroup>
-  );
-});
-
-CanvasModeSwitcher.displayName = 'CanvasModeSwitcher';

invokeai/frontend/web/src/features/controlLayers/components/CanvasPanelContent.tsx

@@ -9,6 +9,7 @@ import { memo } from 'react';
 
 export const CanvasPanelContent = memo(() => {
   const hasEntities = useAppSelector(selectHasEntities);
+
   return (
     <CanvasManagerProviderGate>
       <Flex flexDir="column" gap={2} w="full" h="full">

invokeai/frontend/web/src/features/controlLayers/components/CanvasSendToToggle.tsx

@@ -0,0 +1,65 @@
+import { Flex, Text } from '@invoke-ai/ui-library';
+import { createSelector } from '@reduxjs/toolkit';
+import { useAppDispatch, useAppSelector } from 'app/store/storeHooks';
+import { IconSwitch } from 'common/components/IconSwitch';
+import {
+  selectCanvasSettingsSlice,
+  settingsSendToCanvasChanged,
+} from 'features/controlLayers/store/canvasSettingsSlice';
+import { memo, useCallback } from 'react';
+import { useTranslation } from 'react-i18next';
+import { PiImageBold, PiPaintBrushBold } from 'react-icons/pi';
+
+const TooltipSendToGallery = memo(() => {
+  const { t } = useTranslation();
+
+  return (
+    <Flex flexDir="column">
+      <Text fontWeight="semibold">{t('controlLayers.sendToGallery')}</Text>
+      <Text fontWeight="normal">{t('controlLayers.sendToGalleryDesc')}</Text>
+    </Flex>
+  );
+});
+
+TooltipSendToGallery.displayName = 'TooltipSendToGallery';
+
+const TooltipSendToCanvas = memo(() => {
+  const { t } = useTranslation();
+
+  return (
+    <Flex flexDir="column">
+      <Text fontWeight="semibold">{t('controlLayers.sendToCanvas')}</Text>
+      <Text fontWeight="normal">{t('controlLayers.sendToCanvasDesc')}</Text>
+    </Flex>
+  );
+});
+
+TooltipSendToCanvas.displayName = 'TooltipSendToCanvas';
+
+const selectSendToCanvas = createSelector(selectCanvasSettingsSlice, (canvasSettings) => canvasSettings.sendToCanvas);
+
+export const CanvasSendToToggle = memo(() => {
+  const dispatch = useAppDispatch();
+  const sendToCanvas = useAppSelector(selectSendToCanvas);
+
+  const onChange = useCallback(
+    (isChecked: boolean) => {
+      dispatch(settingsSendToCanvasChanged(isChecked));
+    },
+    [dispatch]
+  );
+
+  return (
+    <IconSwitch
+      isChecked={sendToCanvas}
+      onChange={onChange}
+      iconUnchecked={<PiImageBold />}
+      tooltipUnchecked={<TooltipSendToGallery />}
+      iconChecked={<PiPaintBrushBold />}
+      tooltipChecked={<TooltipSendToCanvas />}
+      ariaLabel="Toggle canvas mode"
+    />
+  );
+});
+
+CanvasSendToToggle.displayName = 'CanvasSendToToggle';

invokeai/frontend/web/src/features/controlLayers/components/ControlLayer/ControlLayer.tsx

@@ -1,14 +1,13 @@
 import { Spacer } from '@invoke-ai/ui-library';
 import { CanvasEntityContainer } from 'features/controlLayers/components/common/CanvasEntityContainer';
-import { CanvasEntityEnabledToggle } from 'features/controlLayers/components/common/CanvasEntityEnabledToggle';
 import { CanvasEntityHeader } from 'features/controlLayers/components/common/CanvasEntityHeader';
-import { CanvasEntityIsLockedToggle } from 'features/controlLayers/components/common/CanvasEntityIsLockedToggle';
+import { CanvasEntityHeaderCommonActions } from 'features/controlLayers/components/common/CanvasEntityHeaderCommonActions';
 import { CanvasEntityPreviewImage } from 'features/controlLayers/components/common/CanvasEntityPreviewImage';
 import { CanvasEntitySettingsWrapper } from 'features/controlLayers/components/common/CanvasEntitySettingsWrapper';
 import { CanvasEntityEditableTitle } from 'features/controlLayers/components/common/CanvasEntityTitleEdit';
 import { ControlLayerBadges } from 'features/controlLayers/components/ControlLayer/ControlLayerBadges';
 import { ControlLayerControlAdapter } from 'features/controlLayers/components/ControlLayer/ControlLayerControlAdapter';
-import { EntityLayerAdapterGate } from 'features/controlLayers/contexts/EntityAdapterContext';
+import { ControlLayerAdapterGate } from 'features/controlLayers/contexts/EntityAdapterContext';
 import { EntityIdentifierContext } from 'features/controlLayers/contexts/EntityIdentifierContext';
 import type { CanvasEntityIdentifier } from 'features/controlLayers/store/types';
 import { memo, useMemo } from 'react';
@@ -22,21 +21,20 @@ export const ControlLayer = memo(({ id }: Props) => {
 
   return (
     <EntityIdentifierContext.Provider value={entityIdentifier}>
-      <EntityLayerAdapterGate>
+      <ControlLayerAdapterGate>
         <CanvasEntityContainer>
           <CanvasEntityHeader>
             <CanvasEntityPreviewImage />
             <CanvasEntityEditableTitle />
             <Spacer />
             <ControlLayerBadges />
-            <CanvasEntityIsLockedToggle />
-            <CanvasEntityEnabledToggle />
+            <CanvasEntityHeaderCommonActions />
           </CanvasEntityHeader>
           <CanvasEntitySettingsWrapper>
             <ControlLayerControlAdapter />
           </CanvasEntitySettingsWrapper>
         </CanvasEntityContainer>
-      </EntityLayerAdapterGate>
+      </ControlLayerAdapterGate>
     </EntityIdentifierContext.Provider>
   );
 });

invokeai/frontend/web/src/features/controlLayers/components/ControlLayer/ControlLayerControlAdapterModel.tsx

@@ -47,7 +47,7 @@ export const ControlLayerControlAdapterModel = memo(({ modelKey, onChange: onCha
         } else {
           canvasManager.filter.$config.set(IMAGE_FILTERS.canny_image_processor.buildDefaults(modelConfig.base));
         }
-        canvasManager.filter.initialize(entityIdentifier);
+        canvasManager.filter.startFilter(entityIdentifier);
         canvasManager.filter.previewFilter();
       }
     },

invokeai/frontend/web/src/features/controlLayers/components/ControlLayer/ControlLayerMenuItemsControlToRaster.tsx

@@ -1,6 +1,7 @@
 import { MenuItem } from '@invoke-ai/ui-library';
 import { useAppDispatch } from 'app/store/storeHooks';
 import { useEntityIdentifierContext } from 'features/controlLayers/contexts/EntityIdentifierContext';
+import { useCanvasIsBusy } from 'features/controlLayers/hooks/useCanvasIsBusy';
 import { controlLayerConvertedToRasterLayer } from 'features/controlLayers/store/canvasSlice';
 import { memo, useCallback } from 'react';
 import { useTranslation } from 'react-i18next';
@@ -9,6 +10,7 @@ import { PiLightningBold } from 'react-icons/pi';
 export const ControlLayerMenuItemsControlToRaster = memo(() => {
   const { t } = useTranslation();
   const dispatch = useAppDispatch();
+  const isBusy = useCanvasIsBusy();
   const entityIdentifier = useEntityIdentifierContext('control_layer');
 
   const convertControlLayerToRasterLayer = useCallback(() => {
@@ -16,7 +18,7 @@ export const ControlLayerMenuItemsControlToRaster = memo(() => {
   }, [dispatch, entityIdentifier]);
 
   return (
-    <MenuItem onClick={convertControlLayerToRasterLayer} icon={<PiLightningBold />}>
+    <MenuItem onClick={convertControlLayerToRasterLayer} icon={<PiLightningBold />} isDisabled={isBusy}>
       {t('controlLayers.convertToRasterLayer')}
     </MenuItem>
   );

invokeai/frontend/web/src/features/controlLayers/components/ControlLayersToolbar.tsx

@@ -1,38 +0,0 @@
-/* eslint-disable i18next/no-literal-string */
-import { Flex, Spacer } from '@invoke-ai/ui-library';
-import { CanvasModeSwitcher } from 'features/controlLayers/components/CanvasModeSwitcher';
-import { CanvasResetViewButton } from 'features/controlLayers/components/CanvasResetViewButton';
-import { CanvasScale } from 'features/controlLayers/components/CanvasScale';
-import { CanvasSettingsPopover } from 'features/controlLayers/components/Settings/CanvasSettingsPopover';
-import { ToolChooser } from 'features/controlLayers/components/Tool/ToolChooser';
-import { ToolFillColorPicker } from 'features/controlLayers/components/Tool/ToolFillColorPicker';
-import { ToolSettings } from 'features/controlLayers/components/Tool/ToolSettings';
-import { UndoRedoButtonGroup } from 'features/controlLayers/components/UndoRedoButtonGroup';
-import { CanvasManagerProviderGate } from 'features/controlLayers/contexts/CanvasManagerProviderGate';
-import { ToggleProgressButton } from 'features/gallery/components/ImageViewer/ToggleProgressButton';
-import { ViewerToggleMenu } from 'features/gallery/components/ImageViewer/ViewerToggleMenu';
-import { memo } from 'react';
-
-export const ControlLayersToolbar = memo(() => {
-  return (
-    <CanvasManagerProviderGate>
-      <Flex w="full" gap={2} alignItems="center">
-        <ToggleProgressButton />
-        <ToolChooser />
-        <Spacer />
-        <ToolSettings />
-        <Spacer />
-        <CanvasScale />
-        <CanvasResetViewButton />
-        <Spacer />
-        <ToolFillColorPicker />
-        <CanvasModeSwitcher />
-        <UndoRedoButtonGroup />
-        <CanvasSettingsPopover />
-        <ViewerToggleMenu />
-      </Flex>
-    </CanvasManagerProviderGate>
-  );
-});
-
-ControlLayersToolbar.displayName = 'ControlLayersToolbar';

invokeai/frontend/web/src/features/controlLayers/components/HeadsUpDisplay.tsx

@@ -1,53 +1,27 @@
-import { Box, Flex, Text } from '@invoke-ai/ui-library';
-import { useStore } from '@nanostores/react';
+import { Grid, GridItem, Text } from '@invoke-ai/ui-library';
 import { createSelector } from '@reduxjs/toolkit';
 import { useAppSelector } from 'app/store/storeHooks';
-import { useCanvasManager } from 'features/controlLayers/contexts/CanvasManagerProviderGate';
 import { selectCanvasSlice } from 'features/controlLayers/store/selectors';
-import { round } from 'lodash-es';
 import { memo } from 'react';
 
 const selectBbox = createSelector(selectCanvasSlice, (canvas) => canvas.bbox);
 
 export const HeadsUpDisplay = memo(() => {
-  const canvasManager = useCanvasManager();
-  const stageAttrs = useStore(canvasManager.stateApi.$stageAttrs);
-  const cursorPos = useStore(canvasManager.stateApi.$lastCursorPos);
-  const isDrawing = useStore(canvasManager.stateApi.$isDrawing);
-  const isMouseDown = useStore(canvasManager.stateApi.$isMouseDown);
-  const lastMouseDownPos = useStore(canvasManager.stateApi.$lastMouseDownPos);
-  const lastAddedPoint = useStore(canvasManager.stateApi.$lastAddedPoint);
   const bbox = useAppSelector(selectBbox);
 
   return (
-    <Flex flexDir="column" bg="blackAlpha.400" borderBottomEndRadius="base" p={2} minW={64} gap={2}>
-      <HUDItem label="Zoom" value={`${round(stageAttrs.scale * 100, 2)}%`} />
-      <HUDItem label="Stage Pos" value={`${round(stageAttrs.x, 3)}, ${round(stageAttrs.y, 3)}`} />
-      <HUDItem
-        label="Stage Size"
-        value={`${round(stageAttrs.width / stageAttrs.scale, 2)}×${round(stageAttrs.height / stageAttrs.scale, 2)} px`}
-      />
-      <HUDItem label="BBox Size" value={`${bbox.rect.width}×${bbox.rect.height} px`} />
-      <HUDItem label="BBox Position" value={`${bbox.rect.x}, ${bbox.rect.y}`} />
-      <HUDItem label="BBox Width % 8" value={round(bbox.rect.width % 8, 2)} />
-      <HUDItem label="BBox Height % 8" value={round(bbox.rect.height % 8, 2)} />
-      <HUDItem label="BBox X % 8" value={round(bbox.rect.x % 8, 2)} />
-      <HUDItem label="BBox Y % 8" value={round(bbox.rect.y % 8, 2)} />
-      <HUDItem
-        label="Cursor Position"
-        value={cursorPos ? `${round(cursorPos.x, 2)}, ${round(cursorPos.y, 2)}` : '?, ?'}
-      />
-      <HUDItem label="Is Drawing" value={isDrawing ? 'True' : 'False'} />
-      <HUDItem label="Is Mouse Down" value={isMouseDown ? 'True' : 'False'} />
-      <HUDItem
-        label="Last Mouse Down Pos"
-        value={lastMouseDownPos ? `${round(lastMouseDownPos.x, 2)}, ${round(lastMouseDownPos.y, 2)}` : '?, ?'}
-      />
-      <HUDItem
-        label="Last Added Point"
-        value={lastAddedPoint ? `${round(lastAddedPoint.x, 2)}, ${round(lastAddedPoint.y, 2)}` : '?, ?'}
-      />
-    </Flex>
+    <Grid
+      bg="base.900"
+      borderBottomEndRadius="base"
+      p={2}
+      gap={2}
+      borderRadius="base"
+      templateColumns="auto auto"
+      opacity={0.6}
+    >
+      <HUDItem label="BBox" value={`${bbox.rect.width}×${bbox.rect.height} px`} />
+      <HUDItem label="Scaled BBox" value={`${bbox.scaledSize.width}×${bbox.scaledSize.height} px`} />
+    </Grid>
   );
 });
 
@@ -55,12 +29,14 @@ HeadsUpDisplay.displayName = 'HeadsUpDisplay';
 
 const HUDItem = memo(({ label, value }: { label: string; value: string | number }) => {
   return (
-    <Box display="inline-block" lineHeight={1}>
-      <Text as="span">{label}: </Text>
-      <Text as="span" fontWeight="semibold">
-        {value}
-      </Text>
-    </Box>
+    <>
+      <GridItem>
+        <Text textAlign="end">{label}: </Text>
+      </GridItem>
+      <GridItem fontWeight="semibold">
+        <Text>{value}</Text>
+      </GridItem>
+    </>
   );
 });
 

invokeai/frontend/web/src/features/controlLayers/components/IPAdapter/IPAdapter.tsx

@@ -1,7 +1,7 @@
 import { Spacer } from '@invoke-ai/ui-library';
 import { CanvasEntityContainer } from 'features/controlLayers/components/common/CanvasEntityContainer';
-import { CanvasEntityEnabledToggle } from 'features/controlLayers/components/common/CanvasEntityEnabledToggle';
 import { CanvasEntityHeader } from 'features/controlLayers/components/common/CanvasEntityHeader';
+import { CanvasEntityHeaderCommonActions } from 'features/controlLayers/components/common/CanvasEntityHeaderCommonActions';
 import { CanvasEntityEditableTitle } from 'features/controlLayers/components/common/CanvasEntityTitleEdit';
 import { IPAdapterSettings } from 'features/controlLayers/components/IPAdapter/IPAdapterSettings';
 import { EntityIdentifierContext } from 'features/controlLayers/contexts/EntityIdentifierContext';
@@ -18,10 +18,10 @@ export const IPAdapter = memo(({ id }: Props) => {
   return (
     <EntityIdentifierContext.Provider value={entityIdentifier}>
       <CanvasEntityContainer>
-        <CanvasEntityHeader ps={4}>
+        <CanvasEntityHeader ps={4} py={5}>
           <CanvasEntityEditableTitle />
           <Spacer />
-          <CanvasEntityEnabledToggle />
+          <CanvasEntityHeaderCommonActions />
         </CanvasEntityHeader>
         <IPAdapterSettings />
       </CanvasEntityContainer>

invokeai/frontend/web/src/features/controlLayers/components/InpaintMask/InpaintMask.tsx

@@ -1,11 +1,10 @@
 import { Spacer } from '@invoke-ai/ui-library';
 import { CanvasEntityContainer } from 'features/controlLayers/components/common/CanvasEntityContainer';
-import { CanvasEntityEnabledToggle } from 'features/controlLayers/components/common/CanvasEntityEnabledToggle';
 import { CanvasEntityHeader } from 'features/controlLayers/components/common/CanvasEntityHeader';
-import { CanvasEntityIsLockedToggle } from 'features/controlLayers/components/common/CanvasEntityIsLockedToggle';
+import { CanvasEntityHeaderCommonActions } from 'features/controlLayers/components/common/CanvasEntityHeaderCommonActions';
 import { CanvasEntityPreviewImage } from 'features/controlLayers/components/common/CanvasEntityPreviewImage';
 import { CanvasEntityEditableTitle } from 'features/controlLayers/components/common/CanvasEntityTitleEdit';
-import { EntityMaskAdapterGate } from 'features/controlLayers/contexts/EntityAdapterContext';
+import { InpaintMaskAdapterGate } from 'features/controlLayers/contexts/EntityAdapterContext';
 import { EntityIdentifierContext } from 'features/controlLayers/contexts/EntityIdentifierContext';
 import type { CanvasEntityIdentifier } from 'features/controlLayers/store/types';
 import { memo, useMemo } from 'react';
@@ -19,17 +18,16 @@ export const InpaintMask = memo(({ id }: Props) => {
 
   return (
     <EntityIdentifierContext.Provider value={entityIdentifier}>
-      <EntityMaskAdapterGate>
+      <InpaintMaskAdapterGate>
         <CanvasEntityContainer>
           <CanvasEntityHeader>
             <CanvasEntityPreviewImage />
             <CanvasEntityEditableTitle />
             <Spacer />
-            <CanvasEntityIsLockedToggle />
-            <CanvasEntityEnabledToggle />
+            <CanvasEntityHeaderCommonActions />
           </CanvasEntityHeader>
         </CanvasEntityContainer>
-      </EntityMaskAdapterGate>
+      </InpaintMaskAdapterGate>
     </EntityIdentifierContext.Provider>
   );
 });

invokeai/frontend/web/src/features/controlLayers/components/RasterLayer/RasterLayer.tsx

@@ -1,11 +1,10 @@
 import { Spacer } from '@invoke-ai/ui-library';
 import { CanvasEntityContainer } from 'features/controlLayers/components/common/CanvasEntityContainer';
-import { CanvasEntityEnabledToggle } from 'features/controlLayers/components/common/CanvasEntityEnabledToggle';
 import { CanvasEntityHeader } from 'features/controlLayers/components/common/CanvasEntityHeader';
-import { CanvasEntityIsLockedToggle } from 'features/controlLayers/components/common/CanvasEntityIsLockedToggle';
+import { CanvasEntityHeaderCommonActions } from 'features/controlLayers/components/common/CanvasEntityHeaderCommonActions';
 import { CanvasEntityPreviewImage } from 'features/controlLayers/components/common/CanvasEntityPreviewImage';
 import { CanvasEntityEditableTitle } from 'features/controlLayers/components/common/CanvasEntityTitleEdit';
-import { EntityLayerAdapterGate } from 'features/controlLayers/contexts/EntityAdapterContext';
+import { RasterLayerAdapterGate } from 'features/controlLayers/contexts/EntityAdapterContext';
 import { EntityIdentifierContext } from 'features/controlLayers/contexts/EntityIdentifierContext';
 import type { CanvasEntityIdentifier } from 'features/controlLayers/store/types';
 import { memo, useMemo } from 'react';
@@ -19,17 +18,16 @@ export const RasterLayer = memo(({ id }: Props) => {
 
   return (
     <EntityIdentifierContext.Provider value={entityIdentifier}>
-      <EntityLayerAdapterGate>
+      <RasterLayerAdapterGate>
         <CanvasEntityContainer>
           <CanvasEntityHeader>
             <CanvasEntityPreviewImage />
             <CanvasEntityEditableTitle />
             <Spacer />
-            <CanvasEntityIsLockedToggle />
-            <CanvasEntityEnabledToggle />
+            <CanvasEntityHeaderCommonActions />
           </CanvasEntityHeader>
         </CanvasEntityContainer>
-      </EntityLayerAdapterGate>
+      </RasterLayerAdapterGate>
     </EntityIdentifierContext.Provider>
   );
 });

invokeai/frontend/web/src/features/controlLayers/components/RasterLayer/RasterLayerMenuItemsRasterToControl.tsx

@@ -1,6 +1,7 @@
 import { MenuItem } from '@invoke-ai/ui-library';
 import { useAppDispatch } from 'app/store/storeHooks';
 import { useEntityIdentifierContext } from 'features/controlLayers/contexts/EntityIdentifierContext';
+import { useCanvasIsBusy } from 'features/controlLayers/hooks/useCanvasIsBusy';
 import { rasterLayerConvertedToControlLayer } from 'features/controlLayers/store/canvasSlice';
 import { memo, useCallback } from 'react';
 import { useTranslation } from 'react-i18next';
@@ -10,13 +11,14 @@ export const RasterLayerMenuItemsRasterToControl = memo(() => {
   const { t } = useTranslation();
   const dispatch = useAppDispatch();
   const entityIdentifier = useEntityIdentifierContext('raster_layer');
+  const isBusy = useCanvasIsBusy();
 
   const convertRasterLayerToControlLayer = useCallback(() => {
     dispatch(rasterLayerConvertedToControlLayer({ entityIdentifier }));
   }, [dispatch, entityIdentifier]);
 
   return (
-    <MenuItem onClick={convertRasterLayerToControlLayer} icon={<PiLightningBold />}>
+    <MenuItem onClick={convertRasterLayerToControlLayer} icon={<PiLightningBold />} isDisabled={isBusy}>
       {t('controlLayers.convertToControlLayer')}
     </MenuItem>
   );

invokeai/frontend/web/src/features/controlLayers/components/RegionalGuidance/RegionalGuidance.tsx

@@ -1,13 +1,12 @@
 import { Spacer } from '@invoke-ai/ui-library';
 import { CanvasEntityContainer } from 'features/controlLayers/components/common/CanvasEntityContainer';
-import { CanvasEntityEnabledToggle } from 'features/controlLayers/components/common/CanvasEntityEnabledToggle';
 import { CanvasEntityHeader } from 'features/controlLayers/components/common/CanvasEntityHeader';
-import { CanvasEntityIsLockedToggle } from 'features/controlLayers/components/common/CanvasEntityIsLockedToggle';
+import { CanvasEntityHeaderCommonActions } from 'features/controlLayers/components/common/CanvasEntityHeaderCommonActions';
 import { CanvasEntityPreviewImage } from 'features/controlLayers/components/common/CanvasEntityPreviewImage';
 import { CanvasEntityEditableTitle } from 'features/controlLayers/components/common/CanvasEntityTitleEdit';
 import { RegionalGuidanceBadges } from 'features/controlLayers/components/RegionalGuidance/RegionalGuidanceBadges';
 import { RegionalGuidanceSettings } from 'features/controlLayers/components/RegionalGuidance/RegionalGuidanceSettings';
-import { EntityMaskAdapterGate } from 'features/controlLayers/contexts/EntityAdapterContext';
+import { RegionalGuidanceAdapterGate } from 'features/controlLayers/contexts/EntityAdapterContext';
 import { EntityIdentifierContext } from 'features/controlLayers/contexts/EntityIdentifierContext';
 import type { CanvasEntityIdentifier } from 'features/controlLayers/store/types';
 import { memo, useMemo } from 'react';
@@ -21,19 +20,18 @@ export const RegionalGuidance = memo(({ id }: Props) => {
 
   return (
     <EntityIdentifierContext.Provider value={entityIdentifier}>
-      <EntityMaskAdapterGate>
+      <RegionalGuidanceAdapterGate>
         <CanvasEntityContainer>
           <CanvasEntityHeader>
             <CanvasEntityPreviewImage />
             <CanvasEntityEditableTitle />
             <Spacer />
             <RegionalGuidanceBadges />
-            <CanvasEntityIsLockedToggle />
-            <CanvasEntityEnabledToggle />
+            <CanvasEntityHeaderCommonActions />
           </CanvasEntityHeader>
           <RegionalGuidanceSettings />
         </CanvasEntityContainer>
-      </EntityMaskAdapterGate>
+      </RegionalGuidanceAdapterGate>
     </EntityIdentifierContext.Provider>
   );
 });

invokeai/frontend/web/src/features/controlLayers/components/RegionalGuidance/RegionalGuidanceMenuItemsAddPromptsAndIPAdapter.tsx

@@ -2,6 +2,7 @@ import { MenuItem } from '@invoke-ai/ui-library';
 import { createMemoizedSelector } from 'app/store/createMemoizedSelector';
 import { useAppDispatch, useAppSelector } from 'app/store/storeHooks';
 import { useEntityIdentifierContext } from 'features/controlLayers/contexts/EntityIdentifierContext';
+import { useCanvasIsBusy } from 'features/controlLayers/hooks/useCanvasIsBusy';
 import {
   rgIPAdapterAdded,
   rgNegativePromptChanged,
@@ -15,6 +16,7 @@ export const RegionalGuidanceMenuItemsAddPromptsAndIPAdapter = memo(() => {
   const entityIdentifier = useEntityIdentifierContext('regional_guidance');
   const { t } = useTranslation();
   const dispatch = useAppDispatch();
+  const isBusy = useCanvasIsBusy();
   const selectValidActions = useMemo(
     () =>
       createMemoizedSelector(selectCanvasSlice, (canvas) => {
@@ -39,13 +41,15 @@ export const RegionalGuidanceMenuItemsAddPromptsAndIPAdapter = memo(() => {
 
   return (
     <>
-      <MenuItem onClick={addPositivePrompt} isDisabled={!validActions.canAddPositivePrompt}>
+      <MenuItem onClick={addPositivePrompt} isDisabled={!validActions.canAddPositivePrompt || isBusy}>
         {t('controlLayers.addPositivePrompt')}
       </MenuItem>
-      <MenuItem onClick={addNegativePrompt} isDisabled={!validActions.canAddNegativePrompt}>
+      <MenuItem onClick={addNegativePrompt} isDisabled={!validActions.canAddNegativePrompt || isBusy}>
         {t('controlLayers.addNegativePrompt')}
       </MenuItem>
-      <MenuItem onClick={addIPAdapter}>{t('controlLayers.addIPAdapter')}</MenuItem>
+      <MenuItem onClick={addIPAdapter} isDisabled={isBusy}>
+        {t('controlLayers.addIPAdapter')}
+      </MenuItem>
     </>
   );
 });

invokeai/frontend/web/src/features/controlLayers/components/Settings/CanvasSettingsClipToBboxCheckbox.tsx

@@ -1,7 +1,7 @@
 import { Checkbox, FormControl, FormLabel } from '@invoke-ai/ui-library';
 import { createSelector } from '@reduxjs/toolkit';
 import { useAppDispatch, useAppSelector } from 'app/store/storeHooks';
-import { clipToBboxChanged, selectCanvasSettingsSlice } from 'features/controlLayers/store/canvasSettingsSlice';
+import { selectCanvasSettingsSlice, settingsClipToBboxChanged } from 'features/controlLayers/store/canvasSettingsSlice';
 import type { ChangeEvent } from 'react';
 import { memo, useCallback } from 'react';
 import { useTranslation } from 'react-i18next';
@@ -13,7 +13,7 @@ export const CanvasSettingsClipToBboxCheckbox = memo(() => {
   const dispatch = useAppDispatch();
   const clipToBbox = useAppSelector(selectClipToBbox);
   const onChange = useCallback(
-    (e: ChangeEvent<HTMLInputElement>) => dispatch(clipToBboxChanged(e.target.checked)),
+    (e: ChangeEvent<HTMLInputElement>) => dispatch(settingsClipToBboxChanged(e.target.checked)),
     [dispatch]
   );
   return (

invokeai/frontend/web/src/features/controlLayers/components/Settings/CanvasSettingsCompositeMaskedRegionsCheckbox.tsx

@@ -0,0 +1,33 @@
+import { Checkbox, FormControl, FormLabel } from '@invoke-ai/ui-library';
+import { createSelector } from '@reduxjs/toolkit';
+import { useAppDispatch, useAppSelector } from 'app/store/storeHooks';
+import {
+  selectCanvasSettingsSlice,
+  settingsCompositeMaskedRegionsChanged,
+} from 'features/controlLayers/store/canvasSettingsSlice';
+import type { ChangeEvent } from 'react';
+import { memo, useCallback } from 'react';
+import { useTranslation } from 'react-i18next';
+
+const selectCompositeMaskedRegions = createSelector(
+  selectCanvasSettingsSlice,
+  (canvasSettings) => canvasSettings.compositeMaskedRegions
+);
+
+export const CanvasSettingsCompositeMaskedRegionsCheckbox = memo(() => {
+  const { t } = useTranslation();
+  const dispatch = useAppDispatch();
+  const compositeMaskedRegions = useAppSelector(selectCompositeMaskedRegions);
+  const onChange = useCallback(
+    (e: ChangeEvent<HTMLInputElement>) => dispatch(settingsCompositeMaskedRegionsChanged(e.target.checked)),
+    [dispatch]
+  );
+  return (
+    <FormControl w="full">
+      <FormLabel flexGrow={1}>{t('controlLayers.compositeMaskedRegions')}</FormLabel>
+      <Checkbox isChecked={compositeMaskedRegions} onChange={onChange} />
+    </FormControl>
+  );
+});
+
+CanvasSettingsCompositeMaskedRegionsCheckbox.displayName = 'CanvasSettingsCompositeMaskedRegionsCheckbox';

invokeai/frontend/web/src/features/controlLayers/components/Settings/CanvasSettingsDynamicGridSwitch.tsx

@@ -1,15 +1,9 @@
 import { FormControl, FormLabel, Switch } from '@invoke-ai/ui-library';
-import { createSelector } from '@reduxjs/toolkit';
 import { useAppDispatch, useAppSelector } from 'app/store/storeHooks';
-import {
-  selectCanvasSettingsSlice,
-  settingsDynamicGridToggled,
-} from 'features/controlLayers/store/canvasSettingsSlice';
+import { selectDynamicGrid, settingsDynamicGridToggled } from 'features/controlLayers/store/canvasSettingsSlice';
 import { memo, useCallback } from 'react';
 import { useTranslation } from 'react-i18next';
 
-const selectDynamicGrid = createSelector(selectCanvasSettingsSlice, (canvasSettings) => canvasSettings.dynamicGrid);
-
 export const CanvasSettingsDynamicGridSwitch = memo(() => {
   const { t } = useTranslation();
   const dispatch = useAppDispatch();

invokeai/frontend/web/src/features/controlLayers/components/Settings/CanvasSettingsInvertScrollCheckbox.tsx

@@ -1,25 +1,33 @@
 import { Checkbox, FormControl, FormLabel } from '@invoke-ai/ui-library';
 import { createSelector } from '@reduxjs/toolkit';
 import { useAppDispatch, useAppSelector } from 'app/store/storeHooks';
-import { invertScrollChanged, selectToolSlice } from 'features/controlLayers/store/toolSlice';
+import {
+  selectCanvasSettingsSlice,
+  settingsInvertScrollForToolWidthChanged,
+} from 'features/controlLayers/store/canvasSettingsSlice';
 import type { ChangeEvent } from 'react';
 import { memo, useCallback } from 'react';
 import { useTranslation } from 'react-i18next';
 
-const selectInvertScroll = createSelector(selectToolSlice, (tool) => tool.invertScroll);
+const selectInvertScrollForToolWidth = createSelector(
+  selectCanvasSettingsSlice,
+  (settings) => settings.invertScrollForToolWidth
+);
 
 export const CanvasSettingsInvertScrollCheckbox = memo(() => {
   const { t } = useTranslation();
   const dispatch = useAppDispatch();
-  const invertScroll = useAppSelector(selectInvertScroll);
+  const invertScrollForToolWidth = useAppSelector(selectInvertScrollForToolWidth);
   const onChange = useCallback(
-    (e: ChangeEvent<HTMLInputElement>) => dispatch(invertScrollChanged(e.target.checked)),
+    (e: ChangeEvent<HTMLInputElement>) => {
+      dispatch(settingsInvertScrollForToolWidthChanged(e.target.checked));
+    },
     [dispatch]
   );
   return (
     <FormControl w="full">
       <FormLabel flexGrow={1}>{t('unifiedCanvas.invertBrushSizeScrollDirection')}</FormLabel>
-      <Checkbox isChecked={invertScroll} onChange={onChange} />
+      <Checkbox isChecked={invertScrollForToolWidth} onChange={onChange} />
     </FormControl>
   );
 });

invokeai/frontend/web/src/features/controlLayers/components/Settings/CanvasSettingsPopover.tsx

@@ -13,11 +13,13 @@ import { CanvasSettingsAutoSaveCheckbox } from 'features/controlLayers/component
 import { CanvasSettingsClearCachesButton } from 'features/controlLayers/components/Settings/CanvasSettingsClearCachesButton';
 import { CanvasSettingsClearHistoryButton } from 'features/controlLayers/components/Settings/CanvasSettingsClearHistoryButton';
 import { CanvasSettingsClipToBboxCheckbox } from 'features/controlLayers/components/Settings/CanvasSettingsClipToBboxCheckbox';
+import { CanvasSettingsCompositeMaskedRegionsCheckbox } from 'features/controlLayers/components/Settings/CanvasSettingsCompositeMaskedRegionsCheckbox';
 import { CanvasSettingsDynamicGridSwitch } from 'features/controlLayers/components/Settings/CanvasSettingsDynamicGridSwitch';
 import { CanvasSettingsInvertScrollCheckbox } from 'features/controlLayers/components/Settings/CanvasSettingsInvertScrollCheckbox';
 import { CanvasSettingsLogDebugInfoButton } from 'features/controlLayers/components/Settings/CanvasSettingsLogDebugInfo';
 import { CanvasSettingsRecalculateRectsButton } from 'features/controlLayers/components/Settings/CanvasSettingsRecalculateRectsButton';
 import { CanvasSettingsResetButton } from 'features/controlLayers/components/Settings/CanvasSettingsResetButton';
+import { CanvasSettingsShowHUDSwitch } from 'features/controlLayers/components/Settings/CanvasSettingsShowHUDSwitch';
 import { memo } from 'react';
 import { useTranslation } from 'react-i18next';
 import { RiSettings4Fill } from 'react-icons/ri';
@@ -36,7 +38,9 @@ export const CanvasSettingsPopover = memo(() => {
             <CanvasSettingsAutoSaveCheckbox />
             <CanvasSettingsInvertScrollCheckbox />
             <CanvasSettingsClipToBboxCheckbox />
+            <CanvasSettingsCompositeMaskedRegionsCheckbox />
             <CanvasSettingsDynamicGridSwitch />
+            <CanvasSettingsShowHUDSwitch />
             <CanvasSettingsResetButton />
             <DebugSettings />
           </Flex>

invokeai/frontend/web/src/features/controlLayers/components/Settings/CanvasSettingsRecalculateRectsButton.tsx

@@ -7,7 +7,7 @@ export const CanvasSettingsRecalculateRectsButton = memo(() => {
   const { t } = useTranslation();
   const canvasManager = useCanvasManager();
   const onClick = useCallback(() => {
-    for (const adapter of canvasManager.adapters.getAll()) {
+    for (const adapter of canvasManager.getAllAdapters()) {
       adapter.transformer.requestRectCalculation();
     }
   }, [canvasManager]);

invokeai/frontend/web/src/features/controlLayers/components/Settings/CanvasSettingsShowHUDSwitch.tsx

@@ -0,0 +1,28 @@
+import { FormControl, FormLabel, Switch } from '@invoke-ai/ui-library';
+import { createSelector } from '@reduxjs/toolkit';
+import { useAppDispatch, useAppSelector } from 'app/store/storeHooks';
+import { selectCanvasSettingsSlice, settingsShowHUDToggled } from 'features/controlLayers/store/canvasSettingsSlice';
+import { memo, useCallback } from 'react';
+import { useTranslation } from 'react-i18next';
+
+const selectShowHUD = createSelector(selectCanvasSettingsSlice, (canvasSettings) => canvasSettings.showHUD);
+
+export const CanvasSettingsShowHUDSwitch = memo(() => {
+  const { t } = useTranslation();
+  const dispatch = useAppDispatch();
+  const showHUD = useAppSelector(selectShowHUD);
+  const onChange = useCallback(() => {
+    dispatch(settingsShowHUDToggled());
+  }, [dispatch]);
+
+  return (
+    <FormControl>
+      <FormLabel m={0} flexGrow={1}>
+        {t('controlLayers.showHUD')}
+      </FormLabel>
+      <Switch size="sm" isChecked={showHUD} onChange={onChange} />
+    </FormControl>
+  );
+});
+
+CanvasSettingsShowHUDSwitch.displayName = 'CanvasSettingsShowHUDSwitch';

invokeai/frontend/web/src/features/controlLayers/components/StageComponent.tsx

@@ -1,27 +1,24 @@
 import { Flex } from '@invoke-ai/ui-library';
 import { useStore } from '@nanostores/react';
-import { createSelector } from '@reduxjs/toolkit';
 import { $socket } from 'app/hooks/useSocketIO';
 import { logger } from 'app/logging/logger';
 import { useAppStore } from 'app/store/nanostores/store';
 import { useAppSelector } from 'app/store/storeHooks';
 import { HeadsUpDisplay } from 'features/controlLayers/components/HeadsUpDisplay';
 import { CanvasManager } from 'features/controlLayers/konva/CanvasManager';
-import { TRANSPARENCY_CHECKER_PATTERN } from 'features/controlLayers/konva/constants';
-import { selectCanvasSettingsSlice } from 'features/controlLayers/store/canvasSettingsSlice';
+import { TRANSPARENCY_CHECKERBOARD_PATTERN_DATAURL } from 'features/controlLayers/konva/patterns/transparency-checkerboard-pattern';
+import { getPrefixedId } from 'features/controlLayers/konva/util';
+import { selectDynamicGrid, selectShowHUD } from 'features/controlLayers/store/canvasSettingsSlice';
 import Konva from 'konva';
 import { memo, useCallback, useEffect, useLayoutEffect, useState } from 'react';
 import { useDevicePixelRatio } from 'use-device-pixel-ratio';
-import { v4 as uuidv4 } from 'uuid';
 
 const log = logger('canvas');
 
-const showHud = false;
-
 // This will log warnings when layers > 5 - maybe use `import.meta.env.MODE === 'development'` instead?
 Konva.showWarnings = false;
 
-const useStageRenderer = (stage: Konva.Stage, container: HTMLDivElement | null, asPreview: boolean) => {
+const useStageRenderer = (stage: Konva.Stage, container: HTMLDivElement | null) => {
   const store = useAppStore();
   const socket = useStore($socket);
   const dpr = useDevicePixelRatio({ round: false });
@@ -42,28 +39,22 @@ const useStageRenderer = (stage: Konva.Stage, container: HTMLDivElement | null,
     const manager = new CanvasManager(stage, container, store, socket);
     manager.initialize();
     return manager.destroy;
-  }, [asPreview, container, socket, stage, store]);
+  }, [container, socket, stage, store]);
 
   useLayoutEffect(() => {
     Konva.pixelRatio = dpr;
   }, [dpr]);
 };
 
-type Props = {
-  asPreview?: boolean;
-};
-
-const selectDynamicGrid = createSelector(selectCanvasSettingsSlice, (canvasSettings) => canvasSettings.dynamicGrid);
-
-export const StageComponent = memo(({ asPreview = false }: Props) => {
+export const StageComponent = memo(() => {
   const dynamicGrid = useAppSelector(selectDynamicGrid);
+  const showHUD = useAppSelector(selectShowHUD);
 
   const [stage] = useState(
     () =>
       new Konva.Stage({
-        id: uuidv4(),
+        id: getPrefixedId('konva_stage'),
         container: document.createElement('div'),
-        listening: !asPreview,
       })
   );
   const [container, setContainer] = useState<HTMLDivElement | null>(null);
@@ -72,7 +63,7 @@ export const StageComponent = memo(({ asPreview = false }: Props) => {
     setContainer(el);
   }, []);
 
-  useStageRenderer(stage, container, asPreview);
+  useStageRenderer(stage, container);
 
   useEffect(
     () => () => {
@@ -87,7 +78,7 @@ export const StageComponent = memo(({ asPreview = false }: Props) => {
         <Flex
           position="absolute"
           borderRadius="base"
-          bgImage={TRANSPARENCY_CHECKER_PATTERN}
+          bgImage={TRANSPARENCY_CHECKERBOARD_PATTERN_DATAURL}
           top={0}
           right={0}
           bottom={0}
@@ -106,9 +97,9 @@ export const StageComponent = memo(({ asPreview = false }: Props) => {
         overflow="hidden"
         data-testid="control-layers-canvas"
       />
-      {!asPreview && (
-        <Flex position="absolute" top={0} insetInlineStart={0} pointerEvents="none">
-          {showHud && <HeadsUpDisplay />}
+      {showHUD && (
+        <Flex position="absolute" top={1} insetInlineStart={1} pointerEvents="none">
+          <HeadsUpDisplay />
         </Flex>
       )}
     </Flex>

invokeai/frontend/web/src/features/controlLayers/components/StagingArea/StagingAreaToolbar.tsx

@@ -45,7 +45,7 @@ export const StagingAreaToolbar = memo(() => {
   const index = useAppSelector(selectStagedImageIndex);
   const selectedImage = useAppSelector(selectSelectedImage);
   const imageCount = useAppSelector(selectImageCount);
-  const shouldShowStagedImage = useStore(canvasManager.stateApi.$shouldShowStagedImage);
+  const shouldShowStagedImage = useStore(canvasManager.stagingArea.$shouldShowStagedImage);
   const isCanvasActive = useStore(INTERACTION_SCOPES.canvas.$isActive);
   const [changeIsImageIntermediate] = useChangeImageIsIntermediateMutation();
   useScopeOnMount('stagingArea');
@@ -83,8 +83,8 @@ export const StagingAreaToolbar = memo(() => {
   }, [dispatch]);
 
   const onToggleShouldShowStagedImage = useCallback(() => {
-    canvasManager.stateApi.$shouldShowStagedImage.set(!shouldShowStagedImage);
-  }, [canvasManager.stateApi.$shouldShowStagedImage, shouldShowStagedImage]);
+    canvasManager.stagingArea.$shouldShowStagedImage.set(!shouldShowStagedImage);
+  }, [canvasManager.stagingArea.$shouldShowStagedImage, shouldShowStagedImage]);
 
   const onSaveStagingImage = useCallback(() => {
     if (!selectedImage) {
@@ -98,9 +98,9 @@ export const StagingAreaToolbar = memo(() => {
     onPrev,
     {
       preventDefault: true,
-      enabled: isCanvasActive,
+      enabled: isCanvasActive && shouldShowStagedImage && imageCount > 1,
     },
-    [isCanvasActive]
+    [isCanvasActive, shouldShowStagedImage, imageCount]
   );
 
   useHotkeys(
@@ -108,9 +108,9 @@ export const StagingAreaToolbar = memo(() => {
     onNext,
     {
       preventDefault: true,
-      enabled: isCanvasActive,
+      enabled: isCanvasActive && shouldShowStagedImage && imageCount > 1,
     },
-    [isCanvasActive]
+    [isCanvasActive, shouldShowStagedImage, imageCount]
   );
 
   useHotkeys(
@@ -118,9 +118,9 @@ export const StagingAreaToolbar = memo(() => {
     onAccept,
     {
       preventDefault: true,
-      enabled: isCanvasActive,
+      enabled: isCanvasActive && shouldShowStagedImage && imageCount > 1,
     },
-    [isCanvasActive]
+    [isCanvasActive, shouldShowStagedImage, imageCount]
   );
 
   const counterText = useMemo(() => {

invokeai/frontend/web/src/features/controlLayers/components/Tool/ToolBboxButton.tsx

@@ -1,10 +1,6 @@
 import { IconButton } from '@invoke-ai/ui-library';
-import { useAppSelector } from 'app/store/storeHooks';
 import { useSelectTool, useToolIsSelected } from 'features/controlLayers/components/Tool/hooks';
-import { useIsFiltering } from 'features/controlLayers/hooks/useIsFiltering';
-import { useIsTransforming } from 'features/controlLayers/hooks/useIsTransforming';
-import { selectIsStaging } from 'features/controlLayers/store/canvasSessionSlice';
-import { memo, useMemo } from 'react';
+import { memo } from 'react';
 import { useHotkeys } from 'react-hotkeys-hook';
 import { useTranslation } from 'react-i18next';
 import { PiBoundingBoxBold } from 'react-icons/pi';
@@ -13,24 +9,18 @@ export const ToolBboxButton = memo(() => {
   const { t } = useTranslation();
   const selectBbox = useSelectTool('bbox');
   const isSelected = useToolIsSelected('bbox');
-  const isFiltering = useIsFiltering();
-  const isTransforming = useIsTransforming();
-  const isStaging = useAppSelector(selectIsStaging);
-  const isDisabled = useMemo(() => {
-    return isTransforming || isFiltering || isStaging;
-  }, [isFiltering, isStaging, isTransforming]);
 
-  useHotkeys('q', selectBbox, { enabled: !isDisabled || isSelected }, [selectBbox, isSelected, isDisabled]);
+  useHotkeys('c', selectBbox, { enabled: !isSelected }, [selectBbox, isSelected]);
 
   return (
     <IconButton
-      aria-label={`${t('controlLayers.tool.bbox')} (Q)`}
-      tooltip={`${t('controlLayers.tool.bbox')} (Q)`}
+      aria-label={`${t('controlLayers.tool.bbox')} (C)`}
+      tooltip={`${t('controlLayers.tool.bbox')} (C)`}
       icon={<PiBoundingBoxBold />}
       colorScheme={isSelected ? 'invokeBlue' : 'base'}
-      variant="outline"
+      variant="solid"
       onClick={selectBbox}
-      isDisabled={isDisabled}
+      isDisabled={isSelected}
     />
   );
 });

invokeai/frontend/web/src/features/controlLayers/components/Tool/ToolBrushButton.tsx

@@ -1,29 +1,16 @@
 import { IconButton } from '@invoke-ai/ui-library';
-import { useAppSelector } from 'app/store/storeHooks';
 import { useSelectTool, useToolIsSelected } from 'features/controlLayers/components/Tool/hooks';
-import { useIsFiltering } from 'features/controlLayers/hooks/useIsFiltering';
-import { useIsTransforming } from 'features/controlLayers/hooks/useIsTransforming';
-import { selectIsStaging } from 'features/controlLayers/store/canvasSessionSlice';
-import { selectIsSelectedEntityDrawable } from 'features/controlLayers/store/selectors';
-import { memo, useMemo } from 'react';
+import { memo } from 'react';
 import { useHotkeys } from 'react-hotkeys-hook';
 import { useTranslation } from 'react-i18next';
 import { PiPaintBrushBold } from 'react-icons/pi';
 
 export const ToolBrushButton = memo(() => {
   const { t } = useTranslation();
-  const isFiltering = useIsFiltering();
-  const isTransforming = useIsTransforming();
-  const isStaging = useAppSelector(selectIsStaging);
-  const selectBrush = useSelectTool('brush');
   const isSelected = useToolIsSelected('brush');
-  const isDrawingToolAllowed = useAppSelector(selectIsSelectedEntityDrawable);
+  const selectBrush = useSelectTool('brush');
 
-  const isDisabled = useMemo(() => {
-    return isTransforming || isFiltering || isStaging || !isDrawingToolAllowed;
-  }, [isDrawingToolAllowed, isFiltering, isStaging, isTransforming]);
-
-  useHotkeys('b', selectBrush, { enabled: !isDisabled || isSelected }, [isDisabled, isSelected, selectBrush]);
+  useHotkeys('b', selectBrush, { enabled: !isSelected }, [isSelected, selectBrush]);
 
   return (
     <IconButton
@@ -31,9 +18,9 @@ export const ToolBrushButton = memo(() => {
       tooltip={`${t('controlLayers.tool.brush')} (B)`}
       icon={<PiPaintBrushBold />}
       colorScheme={isSelected ? 'invokeBlue' : 'base'}
-      variant="outline"
+      variant="solid"
       onClick={selectBrush}
-      isDisabled={isDisabled}
+      isDisabled={isSelected}
     />
   );
 });

invokeai/frontend/web/src/features/controlLayers/components/Tool/ToolBrushWidth.tsx

@@ -1,9 +1,12 @@
 import {
-  CompositeNumberInput,
   CompositeSlider,
   FormControl,
   FormLabel,
+  IconButton,
+  NumberInput,
+  NumberInputField,
   Popover,
+  PopoverAnchor,
   PopoverArrow,
   PopoverBody,
   PopoverContent,
@@ -11,47 +14,172 @@ import {
 } from '@invoke-ai/ui-library';
 import { createSelector } from '@reduxjs/toolkit';
 import { useAppDispatch, useAppSelector } from 'app/store/storeHooks';
-import { brushWidthChanged, selectToolSlice } from 'features/controlLayers/store/toolSlice';
-import { memo, useCallback } from 'react';
+import { useToolIsSelected } from 'features/controlLayers/components/Tool/hooks';
+import { selectCanvasSettingsSlice, settingsBrushWidthChanged } from 'features/controlLayers/store/canvasSettingsSlice';
+import { clamp } from 'lodash-es';
+import type { KeyboardEvent } from 'react';
+import { memo, useCallback, useEffect, useState } from 'react';
+import { useHotkeys } from 'react-hotkeys-hook';
 import { useTranslation } from 'react-i18next';
+import { PiCaretDownBold } from 'react-icons/pi';
 
-const marks = [0, 100, 200, 300];
+const selectBrushWidth = createSelector(selectCanvasSettingsSlice, (settings) => settings.brushWidth);
 const formatPx = (v: number | string) => `${v} px`;
-const selectBrushWidth = createSelector(selectToolSlice, (tool) => tool.brush.width);
+
+function mapSliderValueToRawValue(value: number) {
+  if (value <= 40) {
+    // 0 to 40 on the slider -> 1px to 50px
+    return 1 + (49 * value) / 40;
+  } else if (value <= 70) {
+    // 40 to 70 on the slider -> 50px to 200px
+    return 50 + (150 * (value - 40)) / 30;
+  } else {
+    // 70 to 100 on the slider -> 200px to 600px
+    return 200 + (400 * (value - 70)) / 30;
+  }
+}
+
+function mapRawValueToSliderValue(value: number) {
+  if (value <= 50) {
+    // 1px to 50px -> 0 to 40 on the slider
+    return ((value - 1) * 40) / 49;
+  } else if (value <= 200) {
+    // 50px to 200px -> 40 to 70 on the slider
+    return 40 + ((value - 50) * 30) / 150;
+  } else {
+    // 200px to 600px -> 70 to 100 on the slider
+    return 70 + ((value - 200) * 30) / 400;
+  }
+}
+
+function formatSliderValue(value: number) {
+  return `${String(mapSliderValueToRawValue(value))} px`;
+}
+
+const marks = [
+  mapRawValueToSliderValue(1),
+  mapRawValueToSliderValue(50),
+  mapRawValueToSliderValue(200),
+  mapRawValueToSliderValue(600),
+];
+
+const sliderDefaultValue = mapRawValueToSliderValue(50);
 
 export const ToolBrushWidth = memo(() => {
   const dispatch = useAppDispatch();
   const { t } = useTranslation();
+  const isSelected = useToolIsSelected('brush');
   const width = useAppSelector(selectBrushWidth);
+  const [localValue, setLocalValue] = useState(width);
   const onChange = useCallback(
     (v: number) => {
-      dispatch(brushWidthChanged(Math.round(v)));
+      dispatch(settingsBrushWidthChanged(clamp(Math.round(v), 1, 600)));
     },
     [dispatch]
   );
+
+  const increment = useCallback(() => {
+    let newWidth = Math.round(width * 1.15);
+    if (newWidth === width) {
+      newWidth += 1;
+    }
+    onChange(newWidth);
+  }, [onChange, width]);
+
+  const decrement = useCallback(() => {
+    let newWidth = Math.round(width * 0.85);
+    if (newWidth === width) {
+      newWidth -= 1;
+    }
+    onChange(newWidth);
+  }, [onChange, width]);
+
+  const onChangeSlider = useCallback(
+    (value: number) => {
+      onChange(mapSliderValueToRawValue(value));
+    },
+    [onChange]
+  );
+
+  const onBlur = useCallback(() => {
+    if (isNaN(Number(localValue))) {
+      onChange(50);
+      setLocalValue(50);
+    } else {
+      onChange(localValue);
+    }
+  }, [localValue, onChange]);
+
+  const onChangeNumberInput = useCallback((valueAsString: string, valueAsNumber: number) => {
+    setLocalValue(valueAsNumber);
+  }, []);
+
+  const onKeyDown = useCallback(
+    (e: KeyboardEvent<HTMLInputElement>) => {
+      if (e.key === 'Enter') {
+        onBlur();
+      }
+    },
+    [onBlur]
+  );
+
+  useEffect(() => {
+    setLocalValue(width);
+  }, [width]);
+
+  useHotkeys('[', decrement, { enabled: isSelected }, [decrement, isSelected]);
+  useHotkeys(']', increment, { enabled: isSelected }, [increment, isSelected]);
+
   return (
-    <FormControl w="min-content" gap={2}>
-      <FormLabel m={0}>{t('controlLayers.width')}</FormLabel>
-      <Popover isLazy>
-        <PopoverTrigger>
-          <CompositeNumberInput
+    <Popover>
+      <FormControl w="min-content" gap={2}>
+        <FormLabel m={0}>{t('controlLayers.width')}</FormLabel>
+        <PopoverAnchor>
+          <NumberInput
+            display="flex"
+            alignItems="center"
             min={1}
             max={600}
-            defaultValue={50}
-            value={width}
-            onChange={onChange}
-            w={24}
+            value={localValue}
+            onChange={onChangeNumberInput}
+            onBlur={onBlur}
+            w="76px"
             format={formatPx}
+            defaultValue={50}
+            onKeyDown={onKeyDown}
+            clampValueOnBlur={false}
+          >
+            <NumberInputField paddingInlineEnd={7} />
+            <PopoverTrigger>
+              <IconButton
+                aria-label="open-slider"
+                icon={<PiCaretDownBold />}
+                size="sm"
+                variant="link"
+                position="absolute"
+                insetInlineEnd={0}
+                h="full"
+              />
+            </PopoverTrigger>
+          </NumberInput>
+        </PopoverAnchor>
+      </FormControl>
+      <PopoverContent w={200} pt={0} pb={2} px={4}>
+        <PopoverArrow />
+        <PopoverBody>
+          <CompositeSlider
+            min={0}
+            max={100}
+            value={mapRawValueToSliderValue(localValue)}
+            onChange={onChangeSlider}
+            defaultValue={sliderDefaultValue}
+            marks={marks}
+            formatValue={formatSliderValue}
+            alwaysShowMarks
           />
-        </PopoverTrigger>
-        <PopoverContent w={200} py={2} px={4}>
-          <PopoverArrow />
-          <PopoverBody>
-            <CompositeSlider min={1} max={300} defaultValue={50} value={width} onChange={onChange} marks={marks} />
-          </PopoverBody>
-        </PopoverContent>
-      </Popover>
-    </FormControl>
+        </PopoverBody>
+      </PopoverContent>
+    </Popover>
   );
 });
 

invokeai/frontend/web/src/features/controlLayers/components/Tool/ToolChooser.tsx

@@ -4,16 +4,11 @@ import { ToolBrushButton } from 'features/controlLayers/components/Tool/ToolBrus
 import { ToolColorPickerButton } from 'features/controlLayers/components/Tool/ToolColorPickerButton';
 import { ToolMoveButton } from 'features/controlLayers/components/Tool/ToolMoveButton';
 import { ToolRectButton } from 'features/controlLayers/components/Tool/ToolRectButton';
-import { useCanvasDeleteLayerHotkey } from 'features/controlLayers/hooks/useCanvasDeleteLayerHotkey';
-import { useCanvasResetLayerHotkey } from 'features/controlLayers/hooks/useCanvasResetLayerHotkey';
 
 import { ToolEraserButton } from './ToolEraserButton';
 import { ToolViewButton } from './ToolViewButton';
 
 export const ToolChooser: React.FC = () => {
-  useCanvasResetLayerHotkey();
-  useCanvasDeleteLayerHotkey();
-
   return (
     <>
       <ButtonGroup isAttached>

invokeai/frontend/web/src/features/controlLayers/components/Tool/ToolColorPickerButton.tsx

@@ -1,31 +1,16 @@
 import { IconButton } from '@invoke-ai/ui-library';
-import { useAppSelector } from 'app/store/storeHooks';
 import { useSelectTool, useToolIsSelected } from 'features/controlLayers/components/Tool/hooks';
-import { useIsFiltering } from 'features/controlLayers/hooks/useIsFiltering';
-import { useIsTransforming } from 'features/controlLayers/hooks/useIsTransforming';
-import { selectIsStaging } from 'features/controlLayers/store/canvasSessionSlice';
-import { memo, useMemo } from 'react';
+import { memo } from 'react';
 import { useHotkeys } from 'react-hotkeys-hook';
 import { useTranslation } from 'react-i18next';
 import { PiEyedropperBold } from 'react-icons/pi';
 
 export const ToolColorPickerButton = memo(() => {
   const { t } = useTranslation();
-  const isFiltering = useIsFiltering();
-  const isTransforming = useIsTransforming();
-  const selectColorPicker = useSelectTool('colorPicker');
   const isSelected = useToolIsSelected('colorPicker');
-  const isStaging = useAppSelector(selectIsStaging);
+  const selectColorPicker = useSelectTool('colorPicker');
 
-  const isDisabled = useMemo(() => {
-    return isTransforming || isFiltering || isStaging;
-  }, [isFiltering, isStaging, isTransforming]);
-
-  useHotkeys('i', selectColorPicker, { enabled: !isDisabled || isSelected }, [
-    selectColorPicker,
-    isSelected,
-    isDisabled,
-  ]);
+  useHotkeys('i', selectColorPicker, { enabled: !isSelected }, [selectColorPicker, isSelected]);
 
   return (
     <IconButton
@@ -33,9 +18,9 @@ export const ToolColorPickerButton = memo(() => {
       tooltip={`${t('controlLayers.tool.colorPicker')} (I)`}
       icon={<PiEyedropperBold />}
       colorScheme={isSelected ? 'invokeBlue' : 'base'}
-      variant="outline"
+      variant="solid"
       onClick={selectColorPicker}
-      isDisabled={isDisabled}
+      isDisabled={isSelected}
     />
   );
 });

invokeai/frontend/web/src/features/controlLayers/components/Tool/ToolEraserButton.tsx

@@ -1,28 +1,16 @@
 import { IconButton } from '@invoke-ai/ui-library';
-import { useAppSelector } from 'app/store/storeHooks';
 import { useSelectTool, useToolIsSelected } from 'features/controlLayers/components/Tool/hooks';
-import { useIsFiltering } from 'features/controlLayers/hooks/useIsFiltering';
-import { useIsTransforming } from 'features/controlLayers/hooks/useIsTransforming';
-import { selectIsStaging } from 'features/controlLayers/store/canvasSessionSlice';
-import { selectIsSelectedEntityDrawable } from 'features/controlLayers/store/selectors';
-import { memo, useMemo } from 'react';
+import { memo } from 'react';
 import { useHotkeys } from 'react-hotkeys-hook';
 import { useTranslation } from 'react-i18next';
 import { PiEraserBold } from 'react-icons/pi';
 
 export const ToolEraserButton = memo(() => {
   const { t } = useTranslation();
-  const isFiltering = useIsFiltering();
-  const isTransforming = useIsTransforming();
-  const isStaging = useAppSelector(selectIsStaging);
-  const selectEraser = useSelectTool('eraser');
   const isSelected = useToolIsSelected('eraser');
-  const isDrawingToolAllowed = useAppSelector(selectIsSelectedEntityDrawable);
-  const isDisabled = useMemo(() => {
-    return isTransforming || isFiltering || isStaging || !isDrawingToolAllowed;
-  }, [isDrawingToolAllowed, isFiltering, isStaging, isTransforming]);
+  const selectEraser = useSelectTool('eraser');
 
-  useHotkeys('e', selectEraser, { enabled: !isDisabled || isSelected }, [isDisabled, isSelected, selectEraser]);
+  useHotkeys('e', selectEraser, { enabled: !isSelected }, [isSelected, selectEraser]);
 
   return (
     <IconButton
@@ -30,9 +18,9 @@ export const ToolEraserButton = memo(() => {
       tooltip={`${t('controlLayers.tool.eraser')} (E)`}
       icon={<PiEraserBold />}
       colorScheme={isSelected ? 'invokeBlue' : 'base'}
-      variant="outline"
+      variant="solid"
       onClick={selectEraser}
-      isDisabled={isDisabled}
+      isDisabled={isSelected}
     />
   );
 });

invokeai/frontend/web/src/features/controlLayers/components/Tool/ToolEraserWidth.tsx

@@ -1,9 +1,12 @@
 import {
-  CompositeNumberInput,
   CompositeSlider,
   FormControl,
   FormLabel,
+  IconButton,
+  NumberInput,
+  NumberInputField,
   Popover,
+  PopoverAnchor,
   PopoverArrow,
   PopoverBody,
   PopoverContent,
@@ -11,47 +14,175 @@ import {
 } from '@invoke-ai/ui-library';
 import { createSelector } from '@reduxjs/toolkit';
 import { useAppDispatch, useAppSelector } from 'app/store/storeHooks';
-import { eraserWidthChanged, selectToolSlice } from 'features/controlLayers/store/toolSlice';
-import { memo, useCallback } from 'react';
+import { useToolIsSelected } from 'features/controlLayers/components/Tool/hooks';
+import {
+  selectCanvasSettingsSlice,
+  settingsEraserWidthChanged,
+} from 'features/controlLayers/store/canvasSettingsSlice';
+import { clamp } from 'lodash-es';
+import type { KeyboardEvent } from 'react';
+import { memo, useCallback, useEffect, useState } from 'react';
+import { useHotkeys } from 'react-hotkeys-hook';
 import { useTranslation } from 'react-i18next';
+import { PiCaretDownBold } from 'react-icons/pi';
 
-const marks = [0, 100, 200, 300];
+const selectEraserWidth = createSelector(selectCanvasSettingsSlice, (settings) => settings.eraserWidth);
 const formatPx = (v: number | string) => `${v} px`;
-const selectEraserWidth = createSelector(selectToolSlice, (tool) => tool.eraser.width);
+
+function mapSliderValueToRawValue(value: number) {
+  if (value <= 40) {
+    // 0 to 40 on the slider -> 1px to 50px
+    return 1 + (49 * value) / 40;
+  } else if (value <= 70) {
+    // 40 to 70 on the slider -> 50px to 200px
+    return 50 + (150 * (value - 40)) / 30;
+  } else {
+    // 70 to 100 on the slider -> 200px to 600px
+    return 200 + (400 * (value - 70)) / 30;
+  }
+}
+
+function mapRawValueToSliderValue(value: number) {
+  if (value <= 50) {
+    // 1px to 50px -> 0 to 40 on the slider
+    return ((value - 1) * 40) / 49;
+  } else if (value <= 200) {
+    // 50px to 200px -> 40 to 70 on the slider
+    return 40 + ((value - 50) * 30) / 150;
+  } else {
+    // 200px to 600px -> 70 to 100 on the slider
+    return 70 + ((value - 200) * 30) / 400;
+  }
+}
+
+function formatSliderValue(value: number) {
+  return `${String(mapSliderValueToRawValue(value))} px`;
+}
+
+const marks = [
+  mapRawValueToSliderValue(1),
+  mapRawValueToSliderValue(50),
+  mapRawValueToSliderValue(200),
+  mapRawValueToSliderValue(600),
+];
+
+const sliderDefaultValue = mapRawValueToSliderValue(50);
 
 export const ToolEraserWidth = memo(() => {
   const dispatch = useAppDispatch();
   const { t } = useTranslation();
+  const isSelected = useToolIsSelected('eraser');
   const width = useAppSelector(selectEraserWidth);
+  const [localValue, setLocalValue] = useState(width);
   const onChange = useCallback(
     (v: number) => {
-      dispatch(eraserWidthChanged(Math.round(v)));
+      dispatch(settingsEraserWidthChanged(clamp(Math.round(v), 1, 600)));
     },
     [dispatch]
   );
+
+  const increment = useCallback(() => {
+    let newWidth = Math.round(width * 1.15);
+    if (newWidth === width) {
+      newWidth += 1;
+    }
+    onChange(newWidth);
+  }, [onChange, width]);
+
+  const decrement = useCallback(() => {
+    let newWidth = Math.round(width * 0.85);
+    if (newWidth === width) {
+      newWidth -= 1;
+    }
+    onChange(newWidth);
+  }, [onChange, width]);
+
+  const onChangeSlider = useCallback(
+    (value: number) => {
+      onChange(mapSliderValueToRawValue(value));
+    },
+    [onChange]
+  );
+
+  const onBlur = useCallback(() => {
+    if (isNaN(Number(localValue))) {
+      onChange(50);
+      setLocalValue(50);
+    } else {
+      onChange(localValue);
+    }
+  }, [localValue, onChange]);
+
+  const onChangeNumberInput = useCallback((valueAsString: string, valueAsNumber: number) => {
+    setLocalValue(valueAsNumber);
+  }, []);
+
+  const onKeyDown = useCallback(
+    (e: KeyboardEvent<HTMLInputElement>) => {
+      if (e.key === 'Enter') {
+        onBlur();
+      }
+    },
+    [onBlur]
+  );
+
+  useEffect(() => {
+    setLocalValue(width);
+  }, [width]);
+
+  useHotkeys('[', decrement, { enabled: isSelected }, [decrement, isSelected]);
+  useHotkeys(']', increment, { enabled: isSelected }, [increment, isSelected]);
+
   return (
-    <FormControl w="min-content" gap={2}>
-      <FormLabel m={0}>{t('controlLayers.width')}</FormLabel>
-      <Popover isLazy>
-        <PopoverTrigger>
-          <CompositeNumberInput
+    <Popover>
+      <FormControl w="min-content" gap={2}>
+        <FormLabel m={0}>{t('controlLayers.width')}</FormLabel>
+        <PopoverAnchor>
+          <NumberInput
+            display="flex"
+            alignItems="center"
             min={1}
             max={600}
-            defaultValue={50}
-            value={width}
-            onChange={onChange}
-            w={24}
+            value={localValue}
+            onChange={onChangeNumberInput}
+            onBlur={onBlur}
+            w="76px"
             format={formatPx}
+            defaultValue={50}
+            onKeyDown={onKeyDown}
+            clampValueOnBlur={false}
+          >
+            <NumberInputField paddingInlineEnd={7} />
+            <PopoverTrigger>
+              <IconButton
+                aria-label="open-slider"
+                icon={<PiCaretDownBold />}
+                size="sm"
+                variant="link"
+                position="absolute"
+                insetInlineEnd={0}
+                h="full"
+              />
+            </PopoverTrigger>
+          </NumberInput>
+        </PopoverAnchor>
+      </FormControl>
+      <PopoverContent w={200} pt={0} pb={2} px={4}>
+        <PopoverArrow />
+        <PopoverBody>
+          <CompositeSlider
+            min={0}
+            max={100}
+            value={mapRawValueToSliderValue(localValue)}
+            onChange={onChangeSlider}
+            defaultValue={sliderDefaultValue}
+            marks={marks}
+            formatValue={formatSliderValue}
+            alwaysShowMarks
           />
-        </PopoverTrigger>
-        <PopoverContent w={200} py={2} px={4}>
-          <PopoverArrow />
-          <PopoverBody>
-            <CompositeSlider min={1} max={300} defaultValue={50} value={width} onChange={onChange} marks={marks} />
-          </PopoverBody>
-        </PopoverContent>
-      </Popover>
-    </FormControl>
+        </PopoverBody>
+      </PopoverContent>
+    </Popover>
   );
 });
 

invokeai/frontend/web/src/features/controlLayers/components/Tool/ToolFillColorPicker.tsx

@@ -1,39 +1,35 @@
-import { Flex, Popover, PopoverBody, PopoverContent, PopoverTrigger } from '@invoke-ai/ui-library';
+import { Box, Flex, Popover, PopoverBody, PopoverContent, PopoverTrigger, Tooltip } from '@invoke-ai/ui-library';
 import { createSelector } from '@reduxjs/toolkit';
 import { useAppDispatch, useAppSelector } from 'app/store/storeHooks';
 import IAIColorPicker from 'common/components/IAIColorPicker';
 import { rgbaColorToString } from 'common/util/colorCodeTransformers';
-import { fillChanged, selectToolSlice } from 'features/controlLayers/store/toolSlice';
+import { selectCanvasSettingsSlice, settingsColorChanged } from 'features/controlLayers/store/canvasSettingsSlice';
 import type { RgbaColor } from 'features/controlLayers/store/types';
 import { memo, useCallback } from 'react';
 import { useTranslation } from 'react-i18next';
 
-const selectFill = createSelector(selectToolSlice, (tool) => tool.fill);
+const selectColor = createSelector(selectCanvasSettingsSlice, (settings) => settings.color);
 
-export const ToolFillColorPicker = memo(() => {
+export const ToolColorPicker = memo(() => {
   const { t } = useTranslation();
-  const fill = useAppSelector(selectFill);
+  const fill = useAppSelector(selectColor);
   const dispatch = useAppDispatch();
   const onChange = useCallback(
     (color: RgbaColor) => {
-      dispatch(fillChanged(color));
+      dispatch(settingsColorChanged(color));
     },
     [dispatch]
   );
   return (
     <Popover isLazy>
       <PopoverTrigger>
-        <Flex
-          as="button"
-          aria-label={t('controlLayers.brushColor')}
-          borderRadius="full"
-          borderWidth={1}
-          bg={rgbaColorToString(fill)}
-          w={8}
-          h={8}
-          cursor="pointer"
-          tabIndex={-1}
-        />
+        <Flex role="button" aria-label={t('controlLayers.fill.fillColor')} tabIndex={-1} w={8} h={8}>
+          <Tooltip label={t('controlLayers.fill.fillColor')}>
+            <Flex w="full" h="full" alignItems="center" justifyContent="center">
+              <Box borderRadius="full" w={6} h={6} borderWidth={1} bg={rgbaColorToString(fill)} />
+            </Flex>
+          </Tooltip>
+        </Flex>
       </PopoverTrigger>
       <PopoverContent>
         <PopoverBody minH={64}>
@@ -44,4 +40,4 @@ export const ToolFillColorPicker = memo(() => {
   );
 });
 
-ToolFillColorPicker.displayName = 'ToolFillColorPicker';
+ToolColorPicker.displayName = 'ToolFillColorPicker';

invokeai/frontend/web/src/features/controlLayers/components/Tool/ToolMoveButton.tsx

@@ -1,28 +1,16 @@
 import { IconButton } from '@invoke-ai/ui-library';
-import { useAppSelector } from 'app/store/storeHooks';
 import { useSelectTool, useToolIsSelected } from 'features/controlLayers/components/Tool/hooks';
-import { useIsFiltering } from 'features/controlLayers/hooks/useIsFiltering';
-import { useIsTransforming } from 'features/controlLayers/hooks/useIsTransforming';
-import { selectIsStaging } from 'features/controlLayers/store/canvasSessionSlice';
-import { selectIsSelectedEntityDrawable } from 'features/controlLayers/store/selectors';
-import { memo, useMemo } from 'react';
+import { memo } from 'react';
 import { useHotkeys } from 'react-hotkeys-hook';
 import { useTranslation } from 'react-i18next';
 import { PiCursorBold } from 'react-icons/pi';
 
 export const ToolMoveButton = memo(() => {
   const { t } = useTranslation();
-  const isFiltering = useIsFiltering();
-  const isTransforming = useIsTransforming();
-  const selectMove = useSelectTool('move');
   const isSelected = useToolIsSelected('move');
-  const isStaging = useAppSelector(selectIsStaging);
-  const isDrawingToolAllowed = useAppSelector(selectIsSelectedEntityDrawable);
-  const isDisabled = useMemo(() => {
-    return isTransforming || isFiltering || isStaging || !isDrawingToolAllowed;
-  }, [isDrawingToolAllowed, isFiltering, isStaging, isTransforming]);
+  const selectMove = useSelectTool('move');
 
-  useHotkeys('v', selectMove, { enabled: !isDisabled || isSelected }, [isDisabled, isSelected, selectMove]);
+  useHotkeys('v', selectMove, { enabled: !isSelected }, [isSelected, selectMove]);
 
   return (
     <IconButton
@@ -30,9 +18,9 @@ export const ToolMoveButton = memo(() => {
       tooltip={`${t('controlLayers.tool.move')} (V)`}
       icon={<PiCursorBold />}
       colorScheme={isSelected ? 'invokeBlue' : 'base'}
-      variant="outline"
+      variant="solid"
       onClick={selectMove}
-      isDisabled={isDisabled}
+      isDisabled={isSelected}
     />
   );
 });

invokeai/frontend/web/src/features/controlLayers/components/Tool/ToolRectButton.tsx

@@ -1,29 +1,16 @@
 import { IconButton } from '@invoke-ai/ui-library';
-import { useAppSelector } from 'app/store/storeHooks';
 import { useSelectTool, useToolIsSelected } from 'features/controlLayers/components/Tool/hooks';
-import { useIsFiltering } from 'features/controlLayers/hooks/useIsFiltering';
-import { useIsTransforming } from 'features/controlLayers/hooks/useIsTransforming';
-import { selectIsStaging } from 'features/controlLayers/store/canvasSessionSlice';
-import { selectIsSelectedEntityDrawable } from 'features/controlLayers/store/selectors';
-import { memo, useMemo } from 'react';
+import { memo } from 'react';
 import { useHotkeys } from 'react-hotkeys-hook';
 import { useTranslation } from 'react-i18next';
 import { PiRectangleBold } from 'react-icons/pi';
 
 export const ToolRectButton = memo(() => {
   const { t } = useTranslation();
-  const selectRect = useSelectTool('rect');
   const isSelected = useToolIsSelected('rect');
-  const isFiltering = useIsFiltering();
-  const isTransforming = useIsTransforming();
-  const isStaging = useAppSelector(selectIsStaging);
-  const isDrawingToolAllowed = useAppSelector(selectIsSelectedEntityDrawable);
+  const selectRect = useSelectTool('rect');
 
-  const isDisabled = useMemo(() => {
-    return isTransforming || isFiltering || isStaging || !isDrawingToolAllowed;
-  }, [isDrawingToolAllowed, isFiltering, isStaging, isTransforming]);
-
-  useHotkeys('u', selectRect, { enabled: !isDisabled || isSelected }, [isDisabled, isSelected, selectRect]);
+  useHotkeys('u', selectRect, { enabled: !isSelected }, [isSelected, selectRect]);
 
   return (
     <IconButton
@@ -31,9 +18,9 @@ export const ToolRectButton = memo(() => {
       tooltip={`${t('controlLayers.tool.rectangle')} (U)`}
       icon={<PiRectangleBold />}
       colorScheme={isSelected ? 'invokeBlue' : 'base'}
-      variant="outline"
+      variant="solid"
       onClick={selectRect}
-      isDisabled={isDisabled}
+      isDisabled={isSelected}
     />
   );
 });

invokeai/frontend/web/src/features/controlLayers/components/Tool/ToolSettings.tsx

@@ -6,7 +6,7 @@ import { memo } from 'react';
 
 export const ToolSettings = memo(() => {
   const canvasManager = useCanvasManager();
-  const tool = useStore(canvasManager.stateApi.$tool);
+  const tool = useStore(canvasManager.tool.$tool);
   if (tool === 'brush') {
     return <ToolBrushWidth />;
   }

invokeai/frontend/web/src/features/controlLayers/components/Tool/ToolViewButton.tsx

@@ -1,26 +1,16 @@
 import { IconButton } from '@invoke-ai/ui-library';
-import { useAppSelector } from 'app/store/storeHooks';
 import { useSelectTool, useToolIsSelected } from 'features/controlLayers/components/Tool/hooks';
-import { useIsFiltering } from 'features/controlLayers/hooks/useIsFiltering';
-import { useIsTransforming } from 'features/controlLayers/hooks/useIsTransforming';
-import { selectIsStaging } from 'features/controlLayers/store/canvasSessionSlice';
-import { memo, useMemo } from 'react';
+import { memo } from 'react';
 import { useHotkeys } from 'react-hotkeys-hook';
 import { useTranslation } from 'react-i18next';
 import { PiHandBold } from 'react-icons/pi';
 
 export const ToolViewButton = memo(() => {
   const { t } = useTranslation();
-  const isTransforming = useIsTransforming();
-  const isFiltering = useIsFiltering();
-  const isStaging = useAppSelector(selectIsStaging);
-  const selectView = useSelectTool('view');
   const isSelected = useToolIsSelected('view');
-  const isDisabled = useMemo(() => {
-    return isTransforming || isFiltering || isStaging;
-  }, [isFiltering, isStaging, isTransforming]);
+  const selectView = useSelectTool('view');
 
-  useHotkeys('h', selectView, { enabled: !isDisabled || isSelected }, [selectView, isSelected, isDisabled]);
+  useHotkeys('h', selectView, { enabled: !isSelected }, [selectView, isSelected]);
 
   return (
     <IconButton
@@ -28,9 +18,9 @@ export const ToolViewButton = memo(() => {
       tooltip={`${t('controlLayers.tool.view')} (H)`}
       icon={<PiHandBold />}
       colorScheme={isSelected ? 'invokeBlue' : 'base'}
-      variant="outline"
+      variant="solid"
       onClick={selectView}
-      isDisabled={isDisabled}
+      isDisabled={isSelected}
     />
   );
 });

invokeai/frontend/web/src/features/controlLayers/components/Tool/hooks.ts

@@ -6,14 +6,14 @@ import { useCallback } from 'react';
 
 export const useToolIsSelected = (tool: Tool) => {
   const canvasManager = useCanvasManager();
-  const isSelected = useStore(computed(canvasManager.stateApi.$tool, (t) => t === tool));
+  const isSelected = useStore(computed(canvasManager.tool.$tool, (t) => t === tool));
   return isSelected;
 };
 
 export const useSelectTool = (tool: Tool) => {
   const canvasManager = useCanvasManager();
   const setTool = useCallback(() => {
-    canvasManager.stateApi.$tool.set(tool);
-  }, [canvasManager.stateApi.$tool, tool]);
+    canvasManager.tool.$tool.set(tool);
+  }, [canvasManager.tool.$tool, tool]);
   return setTool;
 };

invokeai/frontend/web/src/features/controlLayers/components/Toolbar/CanvasToolbar.tsx

@@ -0,0 +1,47 @@
+/* eslint-disable i18next/no-literal-string */
+import { Flex, Spacer } from '@invoke-ai/ui-library';
+import { CanvasSettingsPopover } from 'features/controlLayers/components/Settings/CanvasSettingsPopover';
+import { ToolChooser } from 'features/controlLayers/components/Tool/ToolChooser';
+import { ToolColorPicker } from 'features/controlLayers/components/Tool/ToolFillColorPicker';
+import { ToolSettings } from 'features/controlLayers/components/Tool/ToolSettings';
+import { CanvasToolbarResetViewButton } from 'features/controlLayers/components/Toolbar/CanvasToolbarResetViewButton';
+import { CanvasToolbarSaveToGalleryButton } from 'features/controlLayers/components/Toolbar/CanvasToolbarSaveToGalleryButton';
+import { CanvasToolbarScale } from 'features/controlLayers/components/Toolbar/CanvasToolbarScale';
+import { CanvasManagerProviderGate } from 'features/controlLayers/contexts/CanvasManagerProviderGate';
+import { useCanvasDeleteLayerHotkey } from 'features/controlLayers/hooks/useCanvasDeleteLayerHotkey';
+import { useCanvasEntityQuickSwitchHotkey } from 'features/controlLayers/hooks/useCanvasEntityQuickSwitchHotkey';
+import { useCanvasResetLayerHotkey } from 'features/controlLayers/hooks/useCanvasResetLayerHotkey';
+import { useCanvasUndoRedoHotkeys } from 'features/controlLayers/hooks/useCanvasUndoRedoHotkeys';
+import { useNextPrevEntityHotkeys } from 'features/controlLayers/hooks/useNextPrevEntity';
+import { ToggleProgressButton } from 'features/gallery/components/ImageViewer/ToggleProgressButton';
+import { ViewerToggle } from 'features/gallery/components/ImageViewer/ViewerToggleMenu';
+import { memo } from 'react';
+
+export const CanvasToolbar = memo(() => {
+  useCanvasResetLayerHotkey();
+  useCanvasDeleteLayerHotkey();
+  useCanvasUndoRedoHotkeys();
+  useCanvasEntityQuickSwitchHotkey();
+  useNextPrevEntityHotkeys();
+
+  return (
+    <CanvasManagerProviderGate>
+      <Flex w="full" gap={2} alignItems="center">
+        <ToggleProgressButton />
+        <ToolChooser />
+        <Spacer />
+        <ToolSettings />
+        <Spacer />
+        <CanvasToolbarScale />
+        <CanvasToolbarResetViewButton />
+        <Spacer />
+        <ToolColorPicker />
+        <CanvasToolbarSaveToGalleryButton />
+        <CanvasSettingsPopover />
+        <ViewerToggle />
+      </Flex>
+    </CanvasManagerProviderGate>
+  );
+});
+
+CanvasToolbar.displayName = 'CanvasToolbar';

invokeai/frontend/web/src/features/controlLayers/components/Toolbar/CanvasToolbarResetViewButton.tsx

@@ -1,4 +1,4 @@
-import { $shift, IconButton } from '@invoke-ai/ui-library';
+import { $alt, IconButton } from '@invoke-ai/ui-library';
 import { useStore } from '@nanostores/react';
 import { INTERACTION_SCOPES } from 'common/hooks/interactionScopes';
 import { $canvasManager } from 'features/controlLayers/konva/CanvasManager';
@@ -7,7 +7,7 @@ import { useHotkeys } from 'react-hotkeys-hook';
 import { useTranslation } from 'react-i18next';
 import { PiArrowCounterClockwiseBold } from 'react-icons/pi';
 
-export const CanvasResetViewButton = memo(() => {
+export const CanvasToolbarResetViewButton = memo(() => {
   const { t } = useTranslation();
   const canvasManager = useStore($canvasManager);
   const isCanvasActive = useStore(INTERACTION_SCOPES.canvas.$isActive);
@@ -27,7 +27,7 @@ export const CanvasResetViewButton = memo(() => {
   }, [canvasManager]);
 
   const onReset = useCallback(() => {
-    if ($shift.get()) {
+    if ($alt.get()) {
       resetView();
     } else {
       resetZoom();
@@ -35,7 +35,7 @@ export const CanvasResetViewButton = memo(() => {
   }, [resetView, resetZoom]);
 
   useHotkeys('r', resetView, { enabled: isCanvasActive }, [isCanvasActive]);
-  useHotkeys('shift+r', resetZoom, { enabled: isCanvasActive }, [isCanvasActive]);
+  useHotkeys('alt+r', resetZoom, { enabled: isCanvasActive }, [isCanvasActive]);
 
   return (
     <IconButton
@@ -48,4 +48,4 @@ export const CanvasResetViewButton = memo(() => {
   );
 });
 
-CanvasResetViewButton.displayName = 'CanvasResetViewButton';
+CanvasToolbarResetViewButton.displayName = 'CanvasToolbarResetViewButton';

invokeai/frontend/web/src/features/controlLayers/components/Toolbar/CanvasToolbarSaveToGalleryButton.tsx

@@ -0,0 +1,53 @@
+import { IconButton, useShiftModifier } from '@invoke-ai/ui-library';
+import { logger } from 'app/logging/logger';
+import { buildUseBoolean } from 'common/hooks/useBoolean';
+import { isOk, withResultAsync } from 'common/util/result';
+import { useCanvasManager } from 'features/controlLayers/contexts/CanvasManagerProviderGate';
+import { toast } from 'features/toast/toast';
+import { memo, useCallback } from 'react';
+import { useTranslation } from 'react-i18next';
+import { PiFloppyDiskBold } from 'react-icons/pi';
+import { serializeError } from 'serialize-error';
+
+const log = logger('canvas');
+
+const [useIsSaving] = buildUseBoolean(false);
+
+export const CanvasToolbarSaveToGalleryButton = memo(() => {
+  const { t } = useTranslation();
+  const shift = useShiftModifier();
+  const canvasManager = useCanvasManager();
+  const isSaving = useIsSaving();
+
+  const onClick = useCallback(async () => {
+    isSaving.setTrue();
+
+    const rect = shift ? canvasManager.stateApi.getBbox().rect : canvasManager.stage.getVisibleRect('raster_layer');
+
+    const result = await withResultAsync(() =>
+      canvasManager.compositor.rasterizeAndUploadCompositeRasterLayer(rect, true)
+    );
+
+    if (isOk(result)) {
+      toast({ title: t('controlLayers.savedToGalleryOk') });
+    } else {
+      log.error({ error: serializeError(result.error) }, 'Failed to save canvas to gallery');
+      toast({ title: t('controlLayers.savedToGalleryError'), status: 'error' });
+    }
+
+    isSaving.setFalse();
+  }, [canvasManager.compositor, canvasManager.stage, canvasManager.stateApi, isSaving, shift, t]);
+
+  return (
+    <IconButton
+      variant="ghost"
+      onClick={onClick}
+      icon={<PiFloppyDiskBold />}
+      isLoading={isSaving.isTrue}
+      aria-label={shift ? t('controlLayers.saveBboxToGallery') : t('controlLayers.saveCanvasToGallery')}
+      tooltip={shift ? t('controlLayers.saveBboxToGallery') : t('controlLayers.saveCanvasToGallery')}
+    />
+  );
+});
+
+CanvasToolbarSaveToGalleryButton.displayName = 'CanvasToolbarSaveToGalleryButton';

invokeai/frontend/web/src/features/controlLayers/components/Toolbar/CanvasToolbarScale.tsx

@@ -15,9 +15,8 @@ import {
 } from '@invoke-ai/ui-library';
 import { useStore } from '@nanostores/react';
 import { useCanvasManager } from 'features/controlLayers/contexts/CanvasManagerProviderGate';
-import { MAX_CANVAS_SCALE, MIN_CANVAS_SCALE } from 'features/controlLayers/konva/constants';
 import { snapToNearest } from 'features/controlLayers/konva/util';
-import { clamp, round } from 'lodash-es';
+import { round } from 'lodash-es';
 import { computed } from 'nanostores';
 import type { KeyboardEvent } from 'react';
 import { memo, useCallback, useEffect, useState } from 'react';
@@ -32,7 +31,7 @@ function formatPct(v: number | string) {
   return `${round(Number(v), 2).toLocaleString()}%`;
 }
 
-function mapSliderValueToScale(value: number) {
+function mapSliderValueToRawValue(value: number) {
   if (value <= 40) {
     // 0 to 40 -> 10% to 100%
     return 10 + (90 * value) / 40;
@@ -45,64 +44,58 @@ function mapSliderValueToScale(value: number) {
   }
 }
 
-function mapScaleToSliderValue(scale: number) {
-  if (scale <= 100) {
-    return ((scale - 10) * 40) / 90;
-  } else if (scale <= 500) {
-    return 40 + ((scale - 100) * 30) / 400;
+function mapRawValueToSliderValue(value: number) {
+  if (value <= 100) {
+    return ((value - 10) * 40) / 90;
+  } else if (value <= 500) {
+    return 40 + ((value - 100) * 30) / 400;
   } else {
-    return 70 + ((scale - 500) * 30) / 1500;
+    return 70 + ((value - 500) * 30) / 1500;
   }
 }
 
 function formatSliderValue(value: number) {
-  return String(mapSliderValueToScale(value));
+  return String(mapSliderValueToRawValue(value));
 }
 
 const marks = [
-  mapScaleToSliderValue(10),
-  mapScaleToSliderValue(50),
-  mapScaleToSliderValue(100),
-  mapScaleToSliderValue(500),
-  mapScaleToSliderValue(2000),
+  mapRawValueToSliderValue(10),
+  mapRawValueToSliderValue(50),
+  mapRawValueToSliderValue(100),
+  mapRawValueToSliderValue(500),
+  mapRawValueToSliderValue(2000),
 ];
 
-const sliderDefaultValue = mapScaleToSliderValue(100);
+const sliderDefaultValue = mapRawValueToSliderValue(100);
 
 const snapCandidates = marks.slice(1, marks.length - 1);
 
-export const CanvasScale = memo(() => {
+export const CanvasToolbarScale = memo(() => {
   const { t } = useTranslation();
   const canvasManager = useCanvasManager();
-  const scale = useStore(computed(canvasManager.stateApi.$stageAttrs, (attrs) => attrs.scale));
+  const scale = useStore(computed(canvasManager.stage.$stageAttrs, (attrs) => attrs.scale));
   const [localScale, setLocalScale] = useState(scale * 100);
 
   const onChangeSlider = useCallback(
     (scale: number) => {
-      if (!canvasManager) {
-        return;
-      }
       let snappedScale = scale;
       // Do not snap if shift key is held
       if (!$shift.get()) {
         snappedScale = snapToNearest(scale, snapCandidates, 2);
       }
-      const mappedScale = mapSliderValueToScale(snappedScale);
+      const mappedScale = mapSliderValueToRawValue(snappedScale);
       canvasManager.stage.setScale(mappedScale / 100);
     },
     [canvasManager]
   );
 
   const onBlur = useCallback(() => {
-    if (!canvasManager) {
-      return;
-    }
     if (isNaN(Number(localScale))) {
       canvasManager.stage.setScale(1);
       setLocalScale(100);
       return;
     }
-    canvasManager.stage.setScale(clamp(localScale / 100, MIN_CANVAS_SCALE, MAX_CANVAS_SCALE));
+    canvasManager.stage.setScale(localScale / 100);
   }, [canvasManager, localScale]);
 
   const onChangeNumberInput = useCallback((valueAsString: string, valueAsNumber: number) => {
@@ -130,8 +123,8 @@ export const CanvasScale = memo(() => {
           <NumberInput
             display="flex"
             alignItems="center"
-            min={MIN_CANVAS_SCALE * 100}
-            max={MAX_CANVAS_SCALE * 100}
+            min={canvasManager.stage.config.MIN_SCALE * 100}
+            max={canvasManager.stage.config.MAX_SCALE * 100}
             value={localScale}
             onChange={onChangeNumberInput}
             onBlur={onBlur}
@@ -162,7 +155,7 @@ export const CanvasScale = memo(() => {
           <CompositeSlider
             min={0}
             max={100}
-            value={mapScaleToSliderValue(localScale)}
+            value={mapRawValueToSliderValue(localScale)}
             onChange={onChangeSlider}
             defaultValue={sliderDefaultValue}
             marks={marks}
@@ -175,4 +168,4 @@ export const CanvasScale = memo(() => {
   );
 });
 
-CanvasScale.displayName = 'CanvasScale';
+CanvasToolbarScale.displayName = 'CanvasToolbarScale';

invokeai/frontend/web/src/features/controlLayers/components/Transform/Transform.tsx

@@ -1,19 +1,13 @@
 import { Button, ButtonGroup, Flex, Heading, Spacer } from '@invoke-ai/ui-library';
 import { useStore } from '@nanostores/react';
 import { useCanvasManager } from 'features/controlLayers/contexts/CanvasManagerProviderGate';
-import {
-  EntityIdentifierContext,
-  useEntityIdentifierContext,
-} from 'features/controlLayers/contexts/EntityIdentifierContext';
-import { useEntityAdapter } from 'features/controlLayers/hooks/useEntityAdapter';
+import type { CanvasEntityAdapter } from 'features/controlLayers/konva/CanvasEntityAdapter/types';
 import { memo } from 'react';
 import { useTranslation } from 'react-i18next';
-import { PiArrowsCounterClockwiseBold, PiCheckBold, PiXBold } from 'react-icons/pi';
+import { PiArrowsCounterClockwiseBold, PiArrowsOutBold, PiCheckBold, PiXBold } from 'react-icons/pi';
 
-const TransformBox = memo(() => {
+const TransformBox = memo(({ adapter }: { adapter: CanvasEntityAdapter }) => {
   const { t } = useTranslation();
-  const entityIdentifier = useEntityIdentifierContext();
-  const adapter = useEntityAdapter(entityIdentifier);
   const isProcessing = useStore(adapter.transformer.$isProcessing);
 
   return (
@@ -30,9 +24,19 @@ const TransformBox = memo(() => {
       transitionDuration="normal"
     >
       <Heading size="md" color="base.300" userSelect="none">
-        {t('controlLayers.tool.transform')}
+        {t('controlLayers.transform.transform')}
       </Heading>
       <ButtonGroup isAttached={false} size="sm" w="full">
+        <Button
+          leftIcon={<PiArrowsOutBold />}
+          onClick={adapter.transformer.fitProxyRectToBbox}
+          isLoading={isProcessing}
+          loadingText={t('controlLayers.transform.reset')}
+          variant="ghost"
+        >
+          {t('controlLayers.transform.fitToBbox')}
+        </Button>
+        <Spacer />
         <Button
           leftIcon={<PiArrowsCounterClockwiseBold />}
           onClick={adapter.transformer.resetTransform}
@@ -40,9 +44,8 @@ const TransformBox = memo(() => {
           loadingText={t('controlLayers.reset')}
           variant="ghost"
         >
-          {t('accessibility.reset')}
+          {t('controlLayers.transform.reset')}
         </Button>
-        <Spacer />
         <Button
           leftIcon={<PiCheckBold />}
           onClick={adapter.transformer.applyTransform}
@@ -50,7 +53,7 @@ const TransformBox = memo(() => {
           loadingText={t('common.apply')}
           variant="ghost"
         >
-          {t('common.apply')}
+          {t('controlLayers.transform.apply')}
         </Button>
         <Button
           leftIcon={<PiXBold />}
@@ -59,7 +62,7 @@ const TransformBox = memo(() => {
           loadingText={t('common.cancel')}
           variant="ghost"
         >
-          {t('common.cancel')}
+          {t('controlLayers.transform.cancel')}
         </Button>
       </ButtonGroup>
     </Flex>
@@ -70,15 +73,11 @@ TransformBox.displayName = 'Transform';
 
 export const Transform = () => {
   const canvasManager = useCanvasManager();
-  const transformingEntity = useStore(canvasManager.stateApi.$transformingEntity);
+  const adapter = useStore(canvasManager.stateApi.$transformingAdapter);
 
-  if (!transformingEntity) {
+  if (!adapter) {
     return null;
   }
 
-  return (
-    <EntityIdentifierContext.Provider value={transformingEntity}>
-      <TransformBox />
-    </EntityIdentifierContext.Provider>
-  );
+  return <TransformBox adapter={adapter} />;
 };

invokeai/frontend/web/src/features/controlLayers/components/UndoRedoButtonGroup.tsx

@@ -1,53 +0,0 @@
-/* eslint-disable i18next/no-literal-string */
-import { ButtonGroup, IconButton } from '@invoke-ai/ui-library';
-import { useAppSelector } from 'app/store/storeHooks';
-import { canvasRedo, canvasUndo } from 'features/controlLayers/store/canvasSlice';
-import { selectCanvasMayRedo, selectCanvasMayUndo } from 'features/controlLayers/store/selectors';
-import { memo, useCallback } from 'react';
-import { useHotkeys } from 'react-hotkeys-hook';
-import { useTranslation } from 'react-i18next';
-import { PiArrowClockwiseBold, PiArrowCounterClockwiseBold } from 'react-icons/pi';
-import { useDispatch } from 'react-redux';
-
-export const UndoRedoButtonGroup = memo(() => {
-  const { t } = useTranslation();
-  const dispatch = useDispatch();
-
-  const mayUndo = useAppSelector(selectCanvasMayUndo);
-  const handleUndo = useCallback(() => {
-    dispatch(canvasUndo());
-  }, [dispatch]);
-  useHotkeys(['meta+z', 'ctrl+z'], handleUndo, { enabled: mayUndo, preventDefault: true }, [mayUndo, handleUndo]);
-
-  const mayRedo = useAppSelector(selectCanvasMayRedo);
-  const handleRedo = useCallback(() => {
-    dispatch(canvasRedo());
-  }, [dispatch]);
-  useHotkeys(['meta+shift+z', 'ctrl+shift+z'], handleRedo, { enabled: mayRedo, preventDefault: true }, [
-    mayRedo,
-    handleRedo,
-  ]);
-
-  return (
-    <ButtonGroup isAttached={false}>
-      <IconButton
-        aria-label={t('unifiedCanvas.undo')}
-        tooltip={t('unifiedCanvas.undo')}
-        onClick={handleUndo}
-        icon={<PiArrowCounterClockwiseBold />}
-        isDisabled={!mayUndo}
-        variant="ghost"
-      />
-      <IconButton
-        aria-label={t('unifiedCanvas.redo')}
-        tooltip={t('unifiedCanvas.redo')}
-        onClick={handleRedo}
-        icon={<PiArrowClockwiseBold />}
-        isDisabled={!mayRedo}
-        variant="ghost"
-      />
-    </ButtonGroup>
-  );
-});
-
-UndoRedoButtonGroup.displayName = 'UndoRedoButtonGroup';

invokeai/frontend/web/src/features/controlLayers/components/common/CanvasEntityAddOfTypeButton.tsx

@@ -0,0 +1,70 @@
+import { IconButton } from '@invoke-ai/ui-library';
+import { useAppDispatch } from 'app/store/storeHooks';
+import {
+  controlLayerAdded,
+  inpaintMaskAdded,
+  ipaAdded,
+  rasterLayerAdded,
+  rgAdded,
+} from 'features/controlLayers/store/canvasSlice';
+import type { CanvasEntityIdentifier } from 'features/controlLayers/store/types';
+import { memo, useCallback, useMemo } from 'react';
+import { useTranslation } from 'react-i18next';
+import { PiPlusBold } from 'react-icons/pi';
+
+type Props = {
+  type: CanvasEntityIdentifier['type'];
+};
+
+export const CanvasEntityAddOfTypeButton = memo(({ type }: Props) => {
+  const { t } = useTranslation();
+  const dispatch = useAppDispatch();
+  const onClick = useCallback(() => {
+    switch (type) {
+      case 'inpaint_mask':
+        dispatch(inpaintMaskAdded({ isSelected: true }));
+        break;
+      case 'regional_guidance':
+        dispatch(rgAdded({ isSelected: true }));
+        break;
+      case 'raster_layer':
+        dispatch(rasterLayerAdded({ isSelected: true }));
+        break;
+      case 'control_layer':
+        dispatch(controlLayerAdded({ isSelected: true }));
+        break;
+      case 'ip_adapter':
+        dispatch(ipaAdded({ isSelected: true }));
+        break;
+    }
+  }, [dispatch, type]);
+
+  const label = useMemo(() => {
+    switch (type) {
+      case 'inpaint_mask':
+        return t('controlLayers.addInpaintMask');
+      case 'regional_guidance':
+        return t('controlLayers.addRegionalGuidance');
+      case 'raster_layer':
+        return t('controlLayers.addRasterLayer');
+      case 'control_layer':
+        return t('controlLayers.addControlLayer');
+      case 'ip_adapter':
+        return t('controlLayers.addIPAdapter');
+    }
+  }, [type, t]);
+
+  return (
+    <IconButton
+      size="sm"
+      aria-label={label}
+      tooltip={label}
+      variant="link"
+      icon={<PiPlusBold />}
+      onClick={onClick}
+      alignSelf="stretch"
+    />
+  );
+});
+
+CanvasEntityAddOfTypeButton.displayName = 'CanvasEntityAddOfTypeButton';

invokeai/frontend/web/src/features/controlLayers/components/common/CanvasEntityDeleteButton.tsx

@@ -0,0 +1,31 @@
+import { IconButton } from '@invoke-ai/ui-library';
+import { useAppDispatch } from 'app/store/storeHooks';
+import { useEntityIdentifierContext } from 'features/controlLayers/contexts/EntityIdentifierContext';
+import { entityDeleted } from 'features/controlLayers/store/canvasSlice';
+import { memo, useCallback } from 'react';
+import { useTranslation } from 'react-i18next';
+import { PiTrashSimpleFill } from 'react-icons/pi';
+
+export const CanvasEntityDeleteButton = memo(() => {
+  const { t } = useTranslation();
+  const entityIdentifier = useEntityIdentifierContext();
+  const dispatch = useAppDispatch();
+  const onClick = useCallback(() => {
+    dispatch(entityDeleted({ entityIdentifier }));
+  }, [dispatch, entityIdentifier]);
+
+  return (
+    <IconButton
+      size="sm"
+      aria-label={t('common.delete')}
+      tooltip={t('common.delete')}
+      variant="link"
+      alignSelf="stretch"
+      icon={<PiTrashSimpleFill />}
+      onClick={onClick}
+      colorScheme="error"
+    />
+  );
+});
+
+CanvasEntityDeleteButton.displayName = 'CanvasEntityDeleteButton';

invokeai/frontend/web/src/features/controlLayers/components/common/CanvasEntityEnabledToggle.tsx

@@ -21,7 +21,8 @@ export const CanvasEntityEnabledToggle = memo(() => {
       size="sm"
       aria-label={t(isEnabled ? 'common.enabled' : 'common.disabled')}
       tooltip={t(isEnabled ? 'common.enabled' : 'common.disabled')}
-      variant="ghost"
+      variant="link"
+      alignSelf="stretch"
       icon={isEnabled ? <PiCircleFill /> : <PiCircleBold />}
       onClick={onClick}
     />