nodes/kijai-ComfyUI-WanVideoWrapper
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
1,258 Commits 8 Branches 0 Tags
e926f7a069e3efc25bdcc78e1bf65c39ac1a2cd4
Commit Graph
100 Commits
This Branch
This Branch
All Branches
Author SHA1 Message Date
kijai e926f7a069 version bump 1.3.9 2025-11-11 10:57:46 +02:00
kijai e01e34da1f Update nodes_model_loading.py 2025-11-11 10:01:06 +02:00
kijai 47514f678d Allow loading original Ovi -models 2025-11-11 09:46:13 +02:00
kijai de3c9c895a Create wanvideo_1_3B_UniLumos_relight_example_01.json 2025-11-10 19:41:22 +02:00
kijai 4576ddb35e Add node to create input for UniLumos 2025-11-10 19:41:19 +02:00
kijai 68392684b5 Add node to use UniLumos 
Simply allows fore and background latent inputs for UniLumos relight model, example inputs seem to work: https://github.com/alibaba-damo-academy/Lumos-Custom/tree/main/UniLumos/UniLumos/examples
 2025-11-10 19:06:45 +02:00
Jukka Seppänen d3f33a9f09 Update readme.md 2025-11-06 16:38:46 +02:00
kijai d0ef3b5601 Update readme.md 2025-11-06 16:37:50 +02:00
kijai 475f96aede Fix accidental positional arg 2025-11-04 20:53:46 +02:00
kijai 1d0516a2a9 Avoid graph break for LongCat 2025-11-04 10:26:37 +02:00
kijai 8002d8a2f9 This still needed for some reason too 2025-11-04 09:58:40 +02:00
kijai 9a588a42ec Fix some precision issues with unmerged lora 2025-11-04 09:47:38 +02:00
kijai 509d6922f5 Update custom_linear.py 2025-11-04 09:44:07 +02:00
kijai 9fa4140159 Make lora torch.compile optional for unmerged lora application 
This change has caused issues especially with LoRAs that have dynamic rank. Will now be disabled by default, to allow full graph with unmerged LoRAs the option to allow compile is available in the Torch Compile Settings -node
 2025-11-04 01:34:51 +02:00
kijai 8ce6916d72 Fix for some cases of using comfy_chunked rope 2025-11-03 10:36:55 +02:00
kijai 0d0d28569a Fix cases where text encoder isn't used (eg. Minimax remover) 2025-11-03 10:29:10 +02:00
kijai 75109fdb79 Fix custom sigmas with euler 2025-11-03 10:12:05 +02:00
kijai 5eae7087fa Fix for S2V 2025-11-02 01:24:50 +02:00
kijai 393fe78ec2 Update model.py 2025-10-31 23:39:55 +02:00
kijai 5f4020b12d Fix a possible issue with Ovi audio model loading 2025-10-31 17:24:24 +02:00
kijai 5da8a6b169 Fix MultiTalk on some models 2025-10-31 16:50:00 +02:00
kijai ce6e7b501d Fix unmerged LoRA application for certain LoRAs 2025-10-30 23:28:38 +02:00
kijai 366f740d28 Update readme.md 2025-10-30 23:10:06 +02:00
Jukka Seppänen d45fe1ee22 Add note about blocking new accounts from posting issues 
Added a note regarding issue posting restrictions due to bot activity.
 2025-10-30 18:19:45 +02:00
kijai da24890d53 Update pyproject.toml 2025-10-30 17:56:20 +02:00
kijai 95391f403d Update nodes_sampler.py 2025-10-30 17:55:55 +02:00
kijai 9e0b3afe4e version checkpoint 2025-10-30 17:53:29 +02:00
kijai ba1beba982 Create LongCat_TI2V_example_01.json 2025-10-30 17:51:52 +02:00
kijai 64c195167b Update nodes.py 2025-10-30 17:33:42 +02:00
kijai cc9bf1e4f5 Store lora diffs in buffers for GGUF as well 2025-10-30 16:44:03 +02:00
kijai a64f115d35 Fix to previous 2025-10-29 11:06:38 +02:00
kijai e45f6f2fc4 Allow WanVideoScheduler -node to work with the looping samplers 
Was broken for Multitalk/WanAnimate/S2V
 2025-10-29 10:41:33 +02:00
kijai 1cd8df5c00 Update custom_linear.py 2025-10-29 02:50:11 +02:00
kijai d2614a9a49 Merge branch 'main' into longcat 2025-10-29 02:33:37 +02:00
kijai 083a8458c4 Register lora diffs as buffers to allow them to work with block swap 
unmerged loras (non GGUF for now) will now be moved with block swap instead of always loaded from cpu to reduce device transfers and allow torch compile full graph
 2025-10-29 02:33:26 +02:00
kijai 1c2f17e8d7 Add utility node to split sampler from settings 
For cleaner previews
 2025-10-29 02:24:24 +02:00
kijai 9d45b9f0de Use comfy core Conv3D workaround for VAE rather than the fp32 cast 2025-10-29 02:23:49 +02:00
Jukka Seppänen 833c6f50c7 Merge pull request #1581 from chengzeyi/fix-ref-conv-dtype-mismatch 
Fix dtype mismatch in ref_conv forward pass
 2025-10-28 14:27:43 +02:00
kijai eebbcd5ee0 Update model.py 2025-10-28 02:04:50 +02:00
kijai 2633119505 Update custom_linear.py 2025-10-28 01:55:25 +02:00
kijai 90908df260 Update model.py 2025-10-28 01:54:42 +02:00
kijai c80a488f70 Use fp32 norms for other models too and other fixes 2025-10-28 01:52:48 +02:00
kijai e69e068b57 Update gguf.py 2025-10-27 21:02:59 +02:00
kijai 54c45500b0 Apply lora diffs with unmerged loras too 2025-10-27 21:01:45 +02:00
kijai e560366600 Update model.py 2025-10-27 18:55:30 +02:00
kijai f880b321c6 Allow compile in lora application 2025-10-27 01:34:32 +02:00
kijai 51fcbd6b3d Revert "Allow compile here" 
This reverts commit f583b56878.
 2025-10-27 01:07:03 +02:00
kijai f583b56878 Allow compile here 2025-10-27 01:06:39 +02:00
kijai c59e52ca44 Precision adjustments 2025-10-27 00:23:32 +02:00
kijai a0bdf20817 Some cleanup and allow full block swap 2025-10-26 23:05:17 +02:00
kijai 8ad7e50f33 Fix cross attention split point 2025-10-26 22:07:19 +02:00
kijai d504c96174 Separate attention for input images like in original 2025-10-26 19:14:47 +02:00
kijai 43acf83adb Update model.py 2025-10-26 16:57:25 +02:00
kijai fb00932cad Init 
https://huggingface.co/Kijai/LongCat-Video_comfy/tree/main
 2025-10-26 16:36:20 +02:00
kijai d74cfc54e8 Don't zero the extra frames 
Possible fix for SVI "shot" method
 2025-10-24 15:28:35 +03:00
kijai cfa883767f Update utils.py 2025-10-24 12:50:49 +03:00
kijai 88a60d71ab Fix 2025-10-23 23:25:20 +03:00
kijai b3ad381a65 not all torch versions have this 2025-10-23 21:12:45 +03:00
kijai 41168b1e82 Support light VAE 
https://huggingface.co/lightx2v/Autoencoders/tree/main
 2025-10-23 12:31:55 +03:00
kijai 7251c996d2 Load these LoRA keys too 
Even though doesn't seem to do anything? At least silences the errors
 2025-10-22 20:43:45 +03:00
kijai 67fcf0ba52 Reduce needless torch.compile recompiles 2025-10-22 13:29:13 +03:00
kijai aa610cde2b VACE: Support per context window reference images 2025-10-22 11:24:30 +03:00
kijai 6b286552b7 MocHa: remove possibly alpha channel from inputs 2025-10-21 23:26:17 +03:00
kijai 2ea824b9ab Mocha: Fix context window slicing 2025-10-21 20:33:03 +03:00
kijai 001040abc7 Update nodes.py 2025-10-21 20:27:37 +03:00
kijai 746d46e137 WanVideoScheduler: Always show split step sigma 2025-10-21 20:14:54 +03:00
kijai 089329ef2e MoCha: revert some RoPE changes 2025-10-21 20:12:40 +03:00
kijai 5c1f64197e Check for actual scale_weights even if the model for some reason doesn't have the scaled_fp8 key 2025-10-21 19:40:50 +03:00
kijai 54e938cd70 MoCha: Fix context window mask 2025-10-21 19:40:01 +03:00
kijai 7495db7669 Update nodes_sampler.py 2025-10-21 19:15:57 +03:00
kijai 4e36aee658 MoCha: experimental context windows support 2025-10-21 18:32:40 +03:00
kijai 56120d633e Update nodes.py 2025-10-21 17:59:37 +03:00
kijai 1f0861b649 MoCha: modify RoPE function to be more torch.compile friendly 2025-10-21 17:54:25 +03:00
kijai 7916f89c33 Add updated MoCha example for lower VRAM 2025-10-21 17:27:41 +03:00
Jukka Seppänen e294f417c0 Merge pull request #1501 from unrealMJ/mocha 
[Feature] Request to add MoCha: End-to-End Video Character Replacement without Structural Guidance
 2025-10-21 17:17:21 +03:00
Jukka Seppänen 74f33df658 Merge pull request #1493 from HM-RunningHub/fix-uni3c-context-options 
Fix Uni3C + Context Options compatibility (Issue #1491)
 2025-10-20 22:33:38 +03:00
kijai 200f6943e3 Add sageattn mode that allows torch.compile 
Latest wheel from woct0rdho includes the torch.compile fix:

https://github.com/woct0rdho/SageAttention/releases

Based on my quick testing this reduces peak VRAM usage a bit when running sageattn + torch.compile
 2025-10-20 15:16:43 +03:00
kijai 8081e1337c Fix downcasting for fp8_fast and add error to indicate you can't do scaled downcast currently 2025-10-20 13:46:05 +03:00
kijai 9ecb80a58b Support loading VACE LoRAs (for ditto) 2025-10-20 11:49:49 +03:00
kijai 967d15321d fix rcm scheduler check 2025-10-19 19:49:24 +03:00
kijai 9cd79d3d4a Add experimental rCM scheduler 
Based on the original code, works but doesn't feel better than dpm++_sde so far
 2025-10-19 19:34:17 +03:00
kijai 7ecd55e92a bump version 2025-10-19 18:13:09 +03:00
kijai 99c0efdb45 Expose force_parameter_static_shapes in torch.compile options 2025-10-19 18:01:03 +03:00
kijai a721fe3d0f Better (hopefully) FlashVSR frame selection 
First frame was completely discarded before
 2025-10-19 13:49:40 +03:00
kijai e8ecda7240 Update wanvideo_1_3B_FlashVSR_upscale_example.json 2025-10-18 16:16:25 +03:00
kijai 7d88316bf8 Fix for the case of applying unmerged lora to already compiled model 2025-10-18 14:00:18 +03:00
kijai 6bcbfc54cf Fix diffusion forcing 2025-10-17 18:20:45 +03:00
kijai 55463a6290 Fix audio_cfg_scale 2025-10-17 18:12:54 +03:00
kijai 6aa0446706 Ovi: fix decoded output audio shape 
I don't know how it ever worked with any nodes... but now it should work with all audio nodes.
 2025-10-16 18:03:42 +03:00
kijai 848ceb2beb Don't require whole facexlib for lynx 2025-10-16 17:51:21 +03:00
kijai cc06d71ca0 Workaround for bug in pytorch 2.9.0 that makes the VAE use crazy amounts of VRAM 
This is probably caused by:

https://github.com/pytorch/pytorch/pull/164027/files

That disables cudnn when using half precision VAE, this workaround simply uses fp32 for the Conv3D operations.
 2025-10-16 17:10:05 +03:00
kijai e221fdd7ac Fix normal tiny vae 2025-10-16 14:54:00 +03:00
kijai c876108059 Update fp8_e4m3fn compile warning on older arch to indicate it should now work with latest Triton 
https://github.com/woct0rdho/triton-windows/releases/tag/v3.5.0-windows.post21
 2025-10-16 13:08:42 +03:00
kijai 9cae986648 Fix for HuMo/Phantom when using cfg 2025-10-16 13:01:21 +03:00
kijai d2ce150490 Update nodes_sampler.py 2025-10-16 10:49:54 +03:00
kijai 50483a3b54 FlashVSR tweaks 2025-10-16 00:42:33 +03:00
kijai 367c612ddb FlashVSR: Better input frame handling 2025-10-15 23:58:32 +03:00
kijai 9a88b9e40a FlashVSR: Add strength setting 2025-10-15 19:11:41 +03:00
kijai 2fd5bb6ffa Basic context window support for FlashVSR 2025-10-15 18:45:42 +03:00
kijai bb75cddd60 Add minimal FlashVSR upscale support 
https://zhuang2002.github.io/FlashVSR/

This only implements the projection model and the VAE, which seems to be enough for upscaling. This does NOT implement any of the streaming and sparse attention code.
 2025-10-15 18:24:28 +03:00