Jedrzej Kosinski
d1ea8c0a58
Add EasyCache support to AceStep
2025-08-20 22:33:45 -07:00
Jedrzej Kosinski
e07b090845
Fix missing comma, curse you Cursor
2025-08-20 20:51:06 -07:00
Jedrzej Kosinski
3618cb39dc
Merge branch 'master' into easycache
2025-08-20 20:49:53 -07:00
Jedrzej Kosinski
fc892a18cd
Add EasyCache support for Qwen Image
2025-08-20 20:48:18 -07:00
Jedrzej Kosinski
7d9d473193
Make Lumina2 compatible with EasyCache
2025-08-20 20:36:51 -07:00
comfyanonymous
9fa1036f60
Forgot this. ( #9470 )
2025-08-20 23:09:35 -04:00
Jedrzej Kosinski
401757acde
Mark EasyCache/SuperEasyCache as experimental (beta)
2025-08-20 20:08:43 -07:00
saurabh-pingale
0737b7e0d2
fix(userdata): catch invalid workflow filenames ( #9434 ) ( #9445 )
2025-08-20 22:27:57 -04:00
comfyanonymous
0963493a9c
Support for Qwen Diffsynth Controlnets canny and depth. ( #9465 )
...
These are not real controlnets but actually a patch on the model so they
will be treated as such.
Put them in the models/model_patches/ folder.
Use the new ModelPatchLoader and QwenImageDiffsynthControlnet nodes.
2025-08-20 22:26:37 -04:00
Jedrzej Kosinski
66afaaa42b
Change max reuse_threshold to 3.0
2025-08-20 19:16:58 -07:00
Jedrzej Kosinski
e4cc5c4082
Created SuperEasyCache to test what happens if caching and reuse is moved outside the scope of conds, added PREDICT_NOISE wrapper to facilitate this test
2025-08-20 18:35:10 -07:00
Jedrzej Kosinski
136654c48b
Properly consider conds in EasyCache logic
2025-08-20 17:27:14 -07:00
Jedrzej Kosinski
129ad27062
Add subsampling to output_prev (output_prev_subsampled now)
2025-08-20 14:46:54 -07:00
comfyanonymous
e73a9dbe30
Add that qwen edit model is supported to readme. ( #9463 )
2025-08-20 17:34:13 -04:00
Jedrzej Kosinski
7f0380f8a1
Add subsampling for heuristic inputs
2025-08-20 14:28:06 -07:00
Harel Cain
fe01885acf
LTXV: fix key frame noise mask dimensions for when real noise mask exists ( #9425 )
2025-08-20 03:33:10 -04:00
comfyanonymous
7139d6d93f
ComfyUI version 0.3.51
2025-08-20 03:15:30 -04:00
ComfyUI Wiki
2f52e8f05f
Bump template to 0.1.62 ( #9419 )
...
* Bump template to 0.1.61
* Bump template to 0.1.62
2025-08-20 03:15:09 -04:00
Jedrzej Kosinski
0784085cc5
Added DIFFUSION_MODEL wrapper around forward function for wan model
2025-08-20 00:05:26 -07:00
comfyanonymous
8d38ea3bbf
Fix bf16 precision issue with qwen image embeddings. ( #9441 )
2025-08-20 02:58:54 -04:00
Jedrzej Kosinski
54f2ef67bb
Make log statement when not skipping useful, preparing for per-cond caching
2025-08-19 23:40:49 -07:00
comfyanonymous
5a8f502db5
Disable prompt weights for qwen. ( #9438 )
2025-08-20 01:08:11 -04:00
comfyanonymous
7cd2c4bd6a
Qwen rotary embeddings should now match reference code. ( #9437 )
2025-08-20 00:45:27 -04:00
Jedrzej Kosinski
14a3f02d8d
Use sigmas from transformer_options instead of timesteps to be compatible with a greater amount of models, make end_percent work
2025-08-19 21:35:18 -07:00
comfyanonymous
dfa791eb4b
Rope fix for qwen vl. ( #9435 )
2025-08-19 20:47:42 -04:00
Jedrzej Kosinski
6ac21a396c
Fixed math to make threshold work as expected, refactored code to use EasyCacheHolder instead of a dict wrapped by object
2025-08-19 17:01:09 -07:00
comfyanonymous
bddd69618b
Change the TextEncodeQwenImageEdit node to use logic closer to reference. ( #9432 )
2025-08-19 16:49:01 -04:00
Alexander Piskun
54d8fdbed0
feat(api-nodes): add Vidu Video nodes ( #9368 )
2025-08-19 16:30:06 -04:00
Alexander Piskun
d844d8b13b
api_nodes: added release version of google's models ( #9304 )
2025-08-19 16:29:24 -04:00
Alexander Piskun
07a927517c
api_nodes: add GPT-5 series models ( #9325 )
2025-08-19 16:29:01 -04:00
Alexander Piskun
f16a70ba67
api_nodes: add MinimaxHailuoVideoNode node ( #9262 )
2025-08-19 16:28:27 -04:00
Alexander Piskun
36b5127fd3
api_nodes: add kling-v2-1 and v2-1-master ( #9257 )
2025-08-19 16:28:07 -04:00
Jedrzej Kosinski
26d54b1de5
Attempting a universal implementation of EasyCache, starting with flux as test; I screwed up the math a bit, but when I set it just right it works.
2025-08-18 22:01:31 -07:00
comfyanonymous
4977f203fa
P2 of qwen edit model. ( #9412 )
...
* P2 of qwen edit model.
* Typo.
* Fix normal qwen.
* Fix.
* Make the TextEncodeQwenImageEdit also set the ref latent.
If you don't want it to set the ref latent and want to use the
ReferenceLatent node with your custom latent instead just disconnect the
VAE.
2025-08-18 22:38:34 -04:00
Alexander Piskun
bd2ab73976
fix(WAN-nodes): invalid nodeid for WanTrackToVideo ( #9396 )
2025-08-18 03:26:55 -04:00
Christian Byrne
da2efeaec6
Bump frontend to 1.25.9 ( #9394 )
2025-08-17 20:21:02 -07:00
Jedrzej Kosinski
7f3b9b16c6
Make step index detection much more robust ( #9392 )
2025-08-17 18:54:07 -04:00
ComfyUI Wiki
d4e353a94e
Update template to 0.1.60 ( #9377 )
2025-08-17 17:38:40 -04:00
comfyanonymous
ed43784b0d
WIP Qwen edit model: The diffusion model part. ( #9383 )
2025-08-17 16:45:39 -04:00
comfyanonymous
0f2b8525bc
Qwen image model refactor. ( #9375 )
2025-08-16 17:51:28 -04:00
Terry Jia
20a84166d0
record audio node ( #8716 )
...
* record audio node
* sf
2025-08-16 02:07:12 -04:00
Christian Byrne
ed2e33c69a
bump frontend version to 1.25.8 ( #9361 )
2025-08-15 23:32:58 -04:00
comfyanonymous
1702e6df16
Implement wan2.2 camera model. ( #9357 )
...
Use the old WanCameraImageToVideo node.
2025-08-15 17:29:58 -04:00
comfyanonymous
c308a8840a
Add FluxKontextMultiReferenceLatentMethod node. ( #9356 )
...
This node is only useful if someone trains the kontext model to properly
use multiple reference images via the index method.
The default is the offset method which feeds the multiple images like if
they were stitched together as one. This method works with the current
flux kontext model.
2025-08-15 15:50:39 -04:00
Alexander Piskun
027c63f63a
fix(OpenAIGPTImage1): set correct MIME type for multipart uploads to OpenAI edits ( #9348 )
2025-08-15 14:57:47 -04:00
comfyanonymous
e08ecfbd8a
Add warning when using old pytorch. ( #9347 )
2025-08-15 00:22:26 -04:00
comfyanonymous
4e5c230f6a
Fix last commit not working on older pytorch. ( #9346 )
2025-08-14 23:44:02 -04:00
Xiangxi Guo (Ryan)
f0d5d0111f
Avoid torch compile graphbreak for older pytorch versions ( #9344 )
...
Turns out torch.compile has some gaps in context manager decorator
syntax support. I've sent patches to fix that in PyTorch, but it won't
be available for all the folks running older versions of PyTorch, hence
this trivial patch.
2025-08-14 23:41:37 -04:00
comfyanonymous
ad19a069f6
Make SLG nodes work on Qwen Image model. ( #9345 )
2025-08-14 23:16:01 -04:00
Alexander Piskun
5d65d6753b
convert WAN nodes to V3 schema ( #9201 )
2025-08-14 21:48:41 -04:00
