nodes/kijai-ComfyUI-WanVideoWrapper
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
1,234 Commits 8 Branches 0 Tags
da24890d53e80be2273ed2e355850c6e2bbde209
Commit Graph
339 Commits
Author SHA1 Message Date
kijai cc9bf1e4f5 Store lora diffs in buffers for GGUF as well 2025-10-30 16:44:03 +02:00
kijai d2614a9a49 Merge branch 'main' into longcat 2025-10-29 02:33:37 +02:00
kijai 9d45b9f0de Use comfy core Conv3D workaround for VAE rather than the fp32 cast 2025-10-29 02:23:49 +02:00
chengzeyiandClaude d15cf3001f Fix dtype mismatch in ref_conv forward pass 
This commit fixes a RuntimeError that occurs when using Fun-Control
reference images: "Input type (float) and bias type (c10::Half)
should be the same"

Root cause:
- Commit 1ba1a16 changed the dtype handling strategy to convert
  the main latent `x` to `base_dtype` instead of converting
  embeddings to match `x.dtype`
- This caused `fun_ref` input to be in a different dtype than
  the `ref_conv` layer's weights and bias
- Line 2324 already handles this correctly for `attn_cond` by
  converting to `self.attn_conv_in.weight.dtype`

Solution:
- Convert `fun_ref` to match `self.ref_conv.weight.dtype` before
  passing through the convolution layer
- This follows the same pattern used for `attn_cond` on line 2324

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>
 2025-10-28 12:08:23 +00:00
kijai eebbcd5ee0 Update model.py 2025-10-28 02:04:50 +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 e560366600 Update model.py 2025-10-27 18:55:30 +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 41168b1e82 Support light VAE 
https://huggingface.co/lightx2v/Autoencoders/tree/main
 2025-10-23 12:31:55 +03:00
kijai 67fcf0ba52 Reduce needless torch.compile recompiles 2025-10-22 13:29:13 +03:00
kijai 1f0861b649 MoCha: modify RoPE function to be more torch.compile friendly 2025-10-21 17:54:25 +03:00
unrealMJ 88defbfdd1 add MoCha 2025-10-21 09:40:54 +08:00
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 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 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 9a88b9e40a FlashVSR: Add strength setting 2025-10-15 19:11:41 +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
kijai 139bdf827f Squashed commit of the following: 
commit 73dd1a06d33953912f5dd684f168028b14e42a36
Author: kijai <40791699+kijai@users.noreply.github.com>
Date:   Mon Oct 13 19:47:38 2025 +0300

    cleanup

commit 39bc2cecf493e2eb176b55e8841d933f0da1ec39
Author: kijai <40791699+kijai@users.noreply.github.com>
Date:   Mon Oct 13 19:24:20 2025 +0300

    Allow scheduling ovi cfg

commit 2c153c5f324dbd59670ad9c51a7995459504a3cd
Merge: dba7667 32eb6b4
Author: kijai <40791699+kijai@users.noreply.github.com>
Date:   Mon Oct 13 17:48:20 2025 +0300

    Merge branch 'main' into ovi

commit dba76674c71af7bf94c82834a0b0e40d94043c99
Merge: 0f11a43 5a0456e
Author: kijai <40791699+kijai@users.noreply.github.com>
Date:   Sun Oct 12 22:45:43 2025 +0300

    Merge branch 'main' into ovi

commit 0f11a439622799ad8070f8a2b8cc8e6a041b761d
Merge: 0999f50 e2d8c9b
Author: kijai <40791699+kijai@users.noreply.github.com>
Date:   Sat Oct 11 07:48:06 2025 +0300

    Merge branch 'main' into ovi

commit 0999f50cfe025290cd7ce88a8dd1acff0b38d9bd
Merge: d45df1f f1d1c83
Author: kijai <40791699+kijai@users.noreply.github.com>
Date:   Fri Oct 10 22:16:09 2025 +0300

    Merge branch 'main' into ovi

commit d45df1fb5b7c629b15eabc197357d62bdc232aaf
Author: kijai <40791699+kijai@users.noreply.github.com>
Date:   Thu Oct 9 20:21:37 2025 +0300

    Remove dependency for librosa

commit d8e7533fdf7eab1d2489c3e025a908c02d997444
Author: kijai <40791699+kijai@users.noreply.github.com>
Date:   Thu Oct 9 19:57:28 2025 +0300

    Remove omegaconf dependency

commit f4e27ff018e98cb5b09655dceda399baea36b240
Author: kijai <40791699+kijai@users.noreply.github.com>
Date:   Thu Oct 9 19:31:06 2025 +0300

    Fix VACE

commit 35d3df39294831e5e7568b6f7e16d2ecf2d790a0
Author: kijai <40791699+kijai@users.noreply.github.com>
Date:   Thu Oct 9 00:26:40 2025 +0300

    small update

commit 96f8ea1d26869ab7e49e12a07f19d5d5a2023253
Author: kijai <40791699+kijai@users.noreply.github.com>
Date:   Wed Oct 8 22:32:57 2025 +0300

    Create wanvideo_2_2_5B_ovi_testing.json

commit a2511be73b9da7019fd21aeb0b521af941c09150
Author: kijai <40791699+kijai@users.noreply.github.com>
Date:   Wed Oct 8 22:32:54 2025 +0300

    Update nodes_sampler.py

commit d3688b8db71452ea1f7c9a2bc0216441d524e56c
Author: kijai <40791699+kijai@users.noreply.github.com>
Date:   Wed Oct 8 21:43:02 2025 +0300

    Allow EasyCache to work with ovi

commit 586d9148a0306ef5d30e9a971a9c3be4cd3ecc97
Author: kijai <40791699+kijai@users.noreply.github.com>
Date:   Wed Oct 8 19:09:06 2025 +0300

    Update model.py

commit 61eedd2839decdb7d4c2ddd5f1310fdaf49d36ad
Author: kijai <40791699+kijai@users.noreply.github.com>
Date:   Wed Oct 8 19:09:02 2025 +0300

    I2V fix

commit a97fcb1b9ae9fb7bbfdf668c24816e014a1b58d1
Author: kijai <40791699+kijai@users.noreply.github.com>
Date:   Wed Oct 8 17:57:28 2025 +0300

    Add nodes to set audio latent size

commit d41e42a697f3d561dabbc22566f633b5f1bbd952
Author: kijai <40791699+kijai@users.noreply.github.com>
Date:   Wed Oct 8 16:42:04 2025 +0300

    Support loading mmaudio vae from .safetensors

commit 1b0e28ec41e3c97fe1f2f057fef9b9bbcb87bca7
Author: kijai <40791699+kijai@users.noreply.github.com>
Date:   Wed Oct 8 16:19:53 2025 +0300

    Update nodes_sampler.py

commit fbd18f45fe85ede8edcb5aebaea7ceb5b6eab5a2
Author: kijai <40791699+kijai@users.noreply.github.com>
Date:   Wed Oct 8 10:16:44 2025 +0300

    Fixes for other workflows

commit b06993b637198f7fad92208f3b3dc9a7d7f57c7f
Author: kijai <40791699+kijai@users.noreply.github.com>
Date:   Wed Oct 8 09:46:27 2025 +0300

    initial commit

    T2V works
 2025-10-13 20:16:53 +03:00
kijai 32eb6b480d Allow canceling mid step 2025-10-13 17:47:28 +03:00
kijai 6d2ff33466 Fix double encode 2025-10-10 09:36:37 +03:00
kijai d6864ec46e Fix for S2V 2025-10-07 23:48:26 +03:00
kijai 76ea2aa6a6 Update wan_video_vae.py 2025-10-07 11:41:54 +03:00
kijai dd888f7a8d GGUF dtype fix 2025-10-06 22:09:58 +03:00
kijai b27c1e5fcd Fix VACE 2025-10-06 21:29:16 +03:00
kijai bb23263f78 Separate full model's ip and ref layer loading 
Allows using full ref with lite ip adapter, or full ref alone without loading the ip weights
 2025-10-06 18:42:14 +03:00
kijai 174cba5759 Fix some precision issues with some models 2025-10-06 13:20:20 +03:00
kijai 1ad36cc5f9 Merge branch 'main' into lynx 2025-10-02 00:55:48 +03:00
kijai 697fff7442 Add FlowMatchSAODEStableScheduler 
source: https://github.com/eddyhhlure1Eddy/ode-ComfyUI-WanVideoWrapper
 2025-10-02 00:53:49 +03:00
kijai 8360666470 Update model.py 2025-10-01 19:34:17 +03:00
kijai 1316f57390 Update model.py 2025-10-01 19:29:05 +03:00
kijai 0730929558 Allow selecting which blocks to use for lynx ref 2025-10-01 19:04:56 +03:00
kijai 1e88104558 compile fixes and cleanup 2025-10-01 18:44:56 +03:00
kijai 28ae01e11b fixes 2025-10-01 10:05:29 +03:00
kijai 5d3c4223a8 Update model.py 2025-09-30 18:59:14 +03:00
kijai ed167a0135 dtype fixes 2025-09-30 18:33:05 +03:00
kijai 1ba1a1662b cleanup and use fp32 text/time embed if available 2025-09-30 18:10:18 +03:00
kijai aefff3796e uncond 2025-09-28 11:21:05 +03:00
kijai 0ed0f54d11 initial lynx full support 2025-09-28 02:01:01 +03:00
kijai 358baa2c2a Lynx + VACE 2025-09-27 02:05:17 +03:00
kijai ab31158673 init: support lite model 2025-09-27 00:49:02 +03:00
kijai 05474da487 Fix denoise_strength usage 2025-09-26 00:32:13 +03:00
kijai f05865de63 Fix VACE 2025-09-23 19:06:15 +03:00
kijai 840553b261 Update model.py 2025-09-23 19:04:37 +03:00
kijai b8c2bb0a93 Add torch native RMSNorm as a choice 2025-09-23 17:41:24 +03:00