nodes/kijai-ComfyUI-WanVideoWrapper
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
1,315 Commits 8 Branches 0 Tags
2b628669452502780ffe499b2ad5ff389169f4c9
Commit Graph
1315 Commits
This Branch
This Branch
All Branches
Author SHA1 Message Date
kijai 2b62866945 Fix SteadyDancer GGUF dtypes 2025-12-06 17:25:57 +02:00
kijai b5415573c5 Update nodes_model_loading.py 2025-12-05 20:55:33 +02:00
kijai 9e005fab90 Fix scaling for light TAE 2025-12-05 14:03:33 +02:00
kijai f85ea5a48a add "empty_frame_pad_image" input to I2V encode node for easier SVI lora use 
This pads the empty frames with the given image, as should be done with SVI-shot and SVI 2.0 LoRAs
 2025-12-05 13:05:35 +02:00
kijai ee1bb5c5c7 Update nodes_utility.py 2025-12-05 12:07:48 +02:00
kijai 041c31fe2e cleanup 2025-12-05 12:01:34 +02:00
kijai 4e7b8dd92c Clearer error on attention mode imports 2025-12-05 02:00:31 +02:00
kijai e2333d0f04 Update nodes_sampler.py 2025-12-04 16:51:30 +02:00
kijai 2dd4c6bf15 Merge branch 'pr/1705' 2025-12-04 16:48:50 +02:00
kijai 9d0a21228b Update nodes_sampler.py 2025-12-04 16:48:23 +02:00
kijai 001bc0e24a Merge branch 'main' of https://github.com/kijai/ComfyUI-WanVideoWrapper 2025-12-04 16:44:29 +02:00
kijai 7189922f36 Update __init__.py 2025-12-04 16:44:18 +02:00
kijai faf9c5927d Merge branch 'pr/1510' 2025-12-04 16:44:02 +02:00
Jukka Seppänen 6c5dc0ba2a Merge pull request #1546 from wzxysf/main 
Enable vae tiling with end frame
 2025-12-04 16:40:29 +02:00
Jukka Seppänen 995804166d Merge pull request #1651 from jamesjjcondon/patch-1 
Refactor data_mean and data_std initialization
 2025-12-04 16:40:05 +02:00
mossmatrix 4e2bf0f1fa Fix device mismatch error with I2V + Lynx embeds 
Fixed RuntimeError when using I2V embeds with Lynx embeds where tensors
were on different devices (cuda:0 and cpu).

The issue occurred because lynx_ref_text_embed["prompt_embeds"] were not
explicitly moved to the GPU device before being passed to the transformer
during Lynx reference buffer extraction.

Changes:
- Move lynx text embeddings to device in both conditional and unconditional
  buffer extraction calls (lines 1114 and 1128)
- Ensures all tensors are on the same device during cross-attention operations
 2025-12-04 00:34:54 -05:00
kijai b06c7d2d6d Add UltraViCo -sage attention mode and refactor some attention code 
https://github.com/thu-ml/DiT-Extrapolation/
 2025-12-03 13:32:45 +02:00
kijai 014e711972 cleanup 2025-12-03 11:09:09 +02:00
kijai c47b1ded69 Apply conv3d workaround to UniAnimate 2025-12-02 15:00:06 +02:00
kijai 7bc45daaf2 Cleanup whitespaces 2025-12-02 14:50:26 +02:00
kijai aebeeb9160 Code cleanup 2025-12-02 14:49:42 +02:00
kijai e60eb995ee version 1.4.0 2025-12-02 00:46:01 +02:00
kijai b9f6c9aa50 Update __init__.py 2025-12-02 00:45:22 +02:00
kijai c1fbc93521 Add ViBTScheduler to use ViBT models 
https://github.com/Yuanshi9815/ViBT/tree/main
 2025-12-02 00:26:50 +02:00
kijai c4ca252fea Fix compiled LoRA application 
Still needs to be unfused
 2025-12-01 20:38:01 +02:00
kijai c4db00609a Use torch.chunk for chunking 2025-12-01 17:39:50 +02:00
kijai 5a52d6b92f Add VAE feat_cache offloading, VAE tqdm progress bar and memory usage report 2025-12-01 15:05:20 +02:00
kijai a652c55bf7 Remove LoRAs when fully offloading as well 2025-12-01 12:50:22 +02:00
kijai 196d39695f Fix sageattn_varlen 2025-12-01 01:11:06 +02:00
kijai e5be3e5263 Use torch custom_ops to avoid graph breaks with torch.compile 
Hopefully finally fixes the torch.compile VRAM issues...
 2025-12-01 00:29:27 +02:00
kijai a6071c7be5 Merge branch 'main' into steadydancer 2025-11-30 17:56:50 +02:00
kijai 0cba1edd4e Better just not compile this as it's causing issues 2025-11-30 17:53:28 +02:00
kijai a9cd073f29 Remove unnecessary recompile when using cfg 2025-11-30 17:52:56 +02:00
kijai 1e9e2be622 Avoid recompile here 2025-11-30 17:32:24 +02:00
kijai 99c3978da4 Reduce peak VRAM usage when not using torch.compile (and some even with it) 
Found some intermediates that weren't freed which should reduce VRAM usage overall, and modified RoPE application outside torch compile for similar gains than when using torch.compile.
 2025-11-30 17:14:53 +02:00
kijai 30bd7d46eb example update 2025-11-29 02:38:02 +02:00
kijai 0c9d5b8dcc context windows 2025-11-29 02:11:57 +02:00
kijai 66d44ec8db This doesn't really do anything useful 2025-11-28 21:15:25 +02:00
kijai 394c7c13d2 Add strength controls 2025-11-28 21:11:20 +02:00
kijai c9931364b3 Create wanvideo_I2V_steadydancer_testing.json 2025-11-28 20:34:55 +02:00
kijai e54fa5d059 Init 2025-11-28 20:32:16 +02:00
kijai 772642b4f1 Update nodes.py 2025-11-27 20:19:06 +02:00
kijai 472ed70757 Add node to preview image embeds 2025-11-27 20:08:23 +02:00
kijai 44feb24290 Allow TTM to work with 5B models 2025-11-20 11:43:53 +02:00
kijai fa7a967ee7 Fix stand-in 2025-11-17 11:56:02 +02:00
kijai ec161373f4 Update readme.md 2025-11-16 19:30:09 +02:00
kijai 0e3fd0b491 Create wanvideo2_2_I2V_A14B_TimeToMove_example.json 2025-11-16 19:26:51 +02:00
kijai f872460285 Fix TTM for dual sampler setups 2025-11-16 19:23:19 +02:00
kijai b826642a83 Add TTM support (Time To Move) 
https://github.com/time-to-move/TTM
 2025-11-16 17:52:49 +02:00
jamesjjcondon aa9f474958 Refactor data_mean and data_std initialization 
Refactor data_mean and data_std to use register_buffer and remove nn.Buffer.

Tested with pytorch version: 2.4.0+cu121
xformers version: 0.0.27.post2
Set vram state to: LOW_VRAM
Device: cuda:0 NVIDIA GeForce RTX 3090 : cudaMallocAsync
Enabled pinned memory 122281.0
Using xformers attention
Python version: 3.10.12 (main, Aug 15 2025, 14:32:43) [GCC 11.4.0]
ComfyUI version: 0.3.68
ComfyUI frontend version: 1.28.8
 2025-11-14 11:57:26 +10:30