Commit Graph

681 Commits

Author SHA1 Message Date
psychedelicious
b52b271dc4 feat(ui): support string batches 2025-01-17 12:19:04 +11:00
psychedelicious
d88b59c5c4 Revert "feat(ui): rearrange canvas paste back nodes to save an image step"
This reverts commit 7cdda00a54.
2025-01-10 15:59:29 +11:00
Ryan Dick
b272d46056 Enable ability to control the weight of FLUX Control LoRAs. 2024-12-17 13:36:10 +00:00
Ryan Dick
5fcd76a712 Fix frontend FLUX graph construction for FLUX control LoRAs. 2024-12-17 07:28:45 -05:00
Mary Hipp
e0344a302c feat(ui): update FLUX graph building to include control layers with control loras 2024-12-17 07:28:45 -05:00
Brandon Rising
246b59f148 Run pnpm fix, regenerate schema 2024-12-17 07:28:45 -05:00
Brandon Rising
046d19446c Rename Structural Lora to Control Lora 2024-12-17 07:28:45 -05:00
Brandon Rising
f3b253987f Initial setup for flux tools control loras 2024-12-17 07:28:45 -05:00
psychedelicious
5eb382074e tweak(ui): slightly clearer logic for skipping regional guidance 2024-12-02 23:46:21 -05:00
psychedelicious
46aa930526 fix(ui): skip disabled ref images 2024-12-02 23:46:21 -05:00
psychedelicious
08704ee8ec feat(ui): use canvas layer validators in control/ip adapter graph builders 2024-11-29 15:32:48 +10:00
psychedelicious
5910892c33 Merge remote-tracking branch 'origin/main' into ryan/flux-regional-prompting 2024-11-29 15:19:39 +10:00
psychedelicious
df0c7d73f3 feat(ui): use regional guidance validation utils in graph builders 2024-11-29 13:26:09 +10:00
psychedelicious
c276b60af9 tidy(ui): use object for addRegions graph builder util arg 2024-11-29 08:49:41 +10:00
Ryan Dick
fa5653cdf7 Remove unused 'denoise' param to addRegions(). 2024-11-27 17:08:42 +00:00
Ryan Dick
9a7b000995 Update frontend to support regional prompts with FLUX in the canvas. 2024-11-27 17:04:43 +00:00
psychedelicious
ef8ededd2f fix(ui): disable width and height output on image batch output
There's a technical challenge with outputting these values directly. `ImageField` does not store them, so the batch's `ImageField` collection does not have width and height for each image.

In order to set up the batch and pass along width and height for each image, we'd need to make a network request for each image when the user clicks Invoke. It would often be cached, but this will eventually create a scaling issue and poor user experience.

As a very simple workaround, users can output the batch image output into an `Image Primitive` node to access the width and height.

This change is implemented by adding some simple special handling when parsing the output fields for the `image_batch` node.

I'll keep this situation in mind when extending the batching system to other field types.
2024-11-20 11:16:54 -05:00
psychedelicious
bddccf6d2f feat(ui): add graph validation for image collection size 2024-11-18 19:12:27 -08:00
psychedelicious
c813fa3fc0 feat(ui): support min and max length for image collections 2024-11-18 19:12:27 -08:00
psychedelicious
616c0f11e1 feat(ui): image batching in workflows
- Add special handling for `ImageBatchInvocation`
- Add input component for image collections, supporting multi-image upload and dnd
- Minor rework of some hooks for accessing node data
2024-11-18 19:12:27 -08:00
Ryan Dick
e8b202d0a5 Update FLUX IP-Adapter graph construction to optimize for XLabs IP-Adapter v2 over v1. This results in degraded performance with v1 IP-Adapters. 2024-11-18 17:06:53 -08:00
psychedelicious
878093f64e fix(ui): image uploading handling
Rework uploadImage and uploadImages helpers and the RTK listener, ensuring gallery view isn't changed unexpectedly and preventing extraneous toasts.

Fix staging area save to gallery button to essentially make a copy of the image, instead of changing its intermediate status.
2024-11-14 13:38:40 -08:00
Mary Hipp
9179619077 actually use optimized denoising 2024-11-08 20:46:08 -05:00
Mary Hipp
13cb5f0ba2 Merge remote-tracking branch 'origin/main' into ryan/sd3-image-to-image 2024-11-08 20:29:56 -05:00
Ryan Dick
f05efd3270 Fix import for getInfill. 2024-11-08 20:42:44 +00:00
Ryan Dick
7732b5d478 Fix bug related to i2l nodes during graph construction of image-to-image workflows. 2024-11-08 20:15:34 +00:00
Mary Hipp
dff6570078 (ui) SD3 support in linear UI 2024-11-08 13:42:57 -05:00
psychedelicious
a9db2ffefd fix(ui): ensure clip vision model is set correctly for FLUX IP Adapters 2024-11-08 10:02:41 +11:00
psychedelicious
5df3c00e28 feat(ui): remove SerializableObject, use type-fest's JsonObject 2024-11-08 07:39:09 +11:00
Mary Hipp
19f7a1295a (ui): add fields for CLIP-L and CLIP-G, remove MainModelConfig type changes 2024-11-04 12:42:09 -05:00
Brandon Rising
ebabf4f7a8 Setup Model and T5 Encoder selection fields for sd3 nodes 2024-11-04 12:42:09 -05:00
psychedelicious
8f02ce54a0 perf(ui): cache image data & transparency mode during generation mode calculation
Perf boost and reduces the number of images we create on the backend.
2024-10-30 11:05:46 +11:00
psychedelicious
91db9c9300 refactor(ui): generalize compositor methods
`CanvasCompositorModule` had a fairly inflexible API, only supporting compositing all raster layers or inpaint masks.

The API has been generalized work with a list of canvas entities. This enables `Merge Down` and `Merge Selected` functionality (though `Merge Selected` is not part of this set of changes).
2024-10-30 11:05:46 +11:00
Riku
097e92db6a fix(ui): always write seamless metadata
Ensure images without seamless enabled correctly reset the setting
when all parameters are recalled
2024-10-29 20:43:44 +11:00
psychedelicious
7cdda00a54 feat(ui): rearrange canvas paste back nodes to save an image step
We were scaling the unscaled image and mask down before doing the paste-back, but this adds an extraneous step & image output.

We can do the paste-back first, then scale to output size after. So instead of 2 resizes before the paste-back, we have 1 resize after.

The end result is the same.
2024-10-29 11:13:31 +11:00
psychedelicious
a74282bce6 feat(ui): graph builders use objects for arg instead of many args 2024-10-29 11:13:31 +11:00
psychedelicious
107f048c7a feat(ui): extract canvas output node prefix to constant 2024-10-29 11:13:31 +11:00
psychedelicious
3947d4a165 fix(ui): normalize infill alpha to 0-255 when building infill nodes
The browser/UI uses float 0-1 for alpha, while backend uses 0-255. We need to normalize the value when building the infill nodes for outpaint.
2024-10-24 19:22:36 +11:00
psychedelicious
bf3260446d fix(ui): use flux_ip_adapter for flux 2024-10-23 08:30:11 +10:00
Mary Hipp
0a96466b60 feat(ui): add IP adapters to FLUX in linear UI 2024-10-22 15:22:56 -04:00
psychedelicious
7185363f17 fix(ui): edge case where controladapters added counts could be off
We were:
- Incrementing `addedControlNets` or `addedT2IAdapters`
- Attempting to add it, but maybe failing and skipping

Need to swap the order of operations to prevent misreporting of added cnet/t2i.

I don't think this would ever actually cause problems.
2024-10-11 10:37:30 +11:00
Ryan Dick
b1567fe0e4 Make FLUX controlnet node API more like SD API and get it working with linear UI. 2024-10-09 23:38:31 +00:00
Mary Hipp
63a2e17f6b possibly a working FLUX controlnet graph 2024-10-09 15:42:02 -04:00
Ryan Dick
a424552c82 Fix frontend lint errors. 2024-09-30 07:59:14 -04:00
Ryan Dick
3463a968c7 Update Linear UI to support FLUX LoRA models that patch the CLIP model in addition to the transformer. 2024-09-30 07:59:14 -04:00
Ryan Dick
81a4c5c23c Fix prettier error. 2024-09-26 19:54:28 -04:00
Ryan Dick
8efa0668e0 Add exponential scaling of the denoising strength scale for FLUX when optimized inpainting is enabled. 2024-09-26 19:54:28 -04:00
Ryan Dick
fae96f3b9f Remove trajectory_guidance_strength parameter. 2024-09-26 19:54:28 -04:00
Mary Hipp
35685194f3 update default scheduler to dpmpp_3m_k 2024-09-25 17:57:34 +03:00
Ryan Dick
183a67cb1e Merge branch 'main' into ryan/flux-trajectory-guidance 2024-09-20 22:29:34 +00:00