Commit Graph

Select branches

Hide Pull Requests

bindweave

causvid

longcat_avatar

longvie2

main

syncammaster

vap

#1004

#1013

#1019

#1022

#1024

#1024

#1039

#1077

#1077

#1110

#1110

#1122

#1122

#1165

#1199

#1199

#1216

#1240

#1263

#1263

#1313

#1318

#1318

#1321

#136

#1360

#1361

#1386

#1386

#1388

#1388

#1443

#1443

#1444

#1493

#15

#1501

#1510

#1546

#1553

#1566

#1572

#1573

#1581

#1588

#1615

#1615

#1651

#1669

#1669

#1677

#1677

#1696

#1705

#1728

#1728

#1749

#1810

#1810

#1824

#1824

#1840

#1840

#1864

#1864

#1868

#1874

#1874

#1877

#1908

#1908

#249

#266

#275

#275

#353

#354

#358

#364

#38

#397

#420

#5

#537

#537

#545

#612

#632

#642

#643

#657

#658

#671

#685

#687

#706

#706

#768

#799

#828

#844

#844

#862

#876

#883

#907

#926

#94

#952

#955

#975

#975

#980

#982

#994

#994

Mono Color

• 7a0da7708e Merge branch 'main' into vap vap kijai 2025-11-04 23:15:20 +02:00
• a51a53d5b7 Use proper self_attn output layers, and cleanup kijai 2025-11-04 21:13:29 +02:00
• 475f96aede Fix accidental positional arg kijai 2025-11-04 20:53:46 +02:00
• 977f4a5c3a Update model.py kijai 2025-11-04 11:31:08 +02:00
• 2a45675498 Merge branch 'main' into vap kijai 2025-11-04 10:39:35 +02:00
• 1d0516a2a9 Avoid graph break for LongCat kijai 2025-11-04 10:26:37 +02:00
• 8002d8a2f9 This still needed for some reason too kijai 2025-11-04 09:58:40 +02:00
• 9a588a42ec Fix some precision issues with unmerged lora kijai 2025-11-04 09:47:38 +02:00
• 509d6922f5 Update custom_linear.py kijai 2025-11-04 09:44:07 +02:00
• 9fa4140159 Make lora torch.compile optional for unmerged lora application kijai 2025-11-04 01:34:51 +02:00
• 8ce6916d72 Fix for some cases of using comfy_chunked rope kijai 2025-11-03 10:36:55 +02:00
• 0d0d28569a Fix cases where text encoder isn't used (eg. Minimax remover) kijai 2025-11-03 10:29:10 +02:00
• 75109fdb79 Fix custom sigmas with euler kijai 2025-11-03 10:12:05 +02:00
• 5eae7087fa Fix for S2V kijai 2025-11-02 01:24:50 +02:00
• ea414c54ac Create wanvideo_I2V_video-as-prompt_testing_WIP.json kijai 2025-11-01 16:53:15 +02:00
• 0013ae0ece Init VAP kijai 2025-11-01 16:47:37 +02:00
• 393fe78ec2 Update model.py kijai 2025-10-31 23:39:55 +02:00
• 0e904e6035 Remove unnecessary casts kijai 2025-10-31 17:28:12 +02:00
• 5f4020b12d Fix a possible issue with Ovi audio model loading kijai 2025-10-31 17:24:24 +02:00
• 5da8a6b169 Fix MultiTalk on some models kijai 2025-10-31 16:50:00 +02:00
• ce6e7b501d Fix unmerged LoRA application for certain LoRAs kijai 2025-10-30 23:28:38 +02:00
• 366f740d28 Update readme.md kijai 2025-10-30 23:10:06 +02:00
• d45fe1ee22 Add note about blocking new accounts from posting issues Jukka Seppänen 2025-10-30 18:19:45 +02:00
• da24890d53 Update pyproject.toml kijai 2025-10-30 17:56:20 +02:00
• 95391f403d Update nodes_sampler.py kijai 2025-10-30 17:55:55 +02:00
• 9e0b3afe4e version checkpoint kijai 2025-10-30 17:53:29 +02:00
• ba1beba982 Create LongCat_TI2V_example_01.json kijai 2025-10-30 17:51:52 +02:00
• 64c195167b Update nodes.py kijai 2025-10-30 17:33:42 +02:00
• cc9bf1e4f5 Store lora diffs in buffers for GGUF as well kijai 2025-10-30 16:44:03 +02:00
• a64f115d35 Fix to previous kijai 2025-10-29 11:06:38 +02:00
• e45f6f2fc4 Allow WanVideoScheduler -node to work with the looping samplers kijai 2025-10-29 10:41:33 +02:00
• 1cd8df5c00 Update custom_linear.py kijai 2025-10-29 02:50:11 +02:00
• d2614a9a49 Merge branch 'main' into longcat kijai 2025-10-29 02:33:37 +02:00
• 083a8458c4 Register lora diffs as buffers to allow them to work with block swap kijai 2025-10-29 02:33:26 +02:00
• 1c2f17e8d7 Add utility node to split sampler from settings kijai 2025-10-29 02:24:24 +02:00
• 9d45b9f0de Use comfy core Conv3D workaround for VAE rather than the fp32 cast kijai 2025-10-29 02:23:49 +02:00
• 833c6f50c7 Merge pull request #1581 from chengzeyi/fix-ref-conv-dtype-mismatch Jukka Seppänen 2025-10-28 14:27:43 +02:00
• d15cf3001f Fix dtype mismatch in ref_conv forward pass chengzeyi 2025-10-28 12:08:23 +00:00
• eebbcd5ee0 Update model.py kijai 2025-10-28 02:04:50 +02:00
• 2633119505 Update custom_linear.py kijai 2025-10-28 01:55:25 +02:00
• 90908df260 Update model.py kijai 2025-10-28 01:54:42 +02:00
• c80a488f70 Use fp32 norms for other models too and other fixes kijai 2025-10-28 01:52:48 +02:00
• e69e068b57 Update gguf.py kijai 2025-10-27 21:02:59 +02:00
• 54c45500b0 Apply lora diffs with unmerged loras too kijai 2025-10-27 21:01:45 +02:00
• e560366600 Update model.py kijai 2025-10-27 18:55:30 +02:00
• f880b321c6 Allow compile in lora application kijai 2025-10-27 01:34:32 +02:00
• 51fcbd6b3d Revert "Allow compile here" kijai 2025-10-27 01:07:03 +02:00
• f583b56878 Allow compile here kijai 2025-10-27 01:06:39 +02:00
• c59e52ca44 Precision adjustments kijai 2025-10-27 00:23:32 +02:00
• a0bdf20817 Some cleanup and allow full block swap kijai 2025-10-26 23:05:17 +02:00
• 8ad7e50f33 Fix cross attention split point kijai 2025-10-26 22:07:19 +02:00
• d504c96174 Separate attention for input images like in original kijai 2025-10-26 19:14:47 +02:00
• 43acf83adb Update model.py kijai 2025-10-26 16:57:25 +02:00
• fb00932cad Init kijai 2025-10-26 16:36:20 +02:00
• 6a37c0b2d6 Enable vae tiling with end frame wzxysf 2025-10-24 22:20:56 +08:00
• d74cfc54e8 Don't zero the extra frames kijai 2025-10-24 15:28:35 +03:00
• cfa883767f Update utils.py kijai 2025-10-24 12:50:49 +03:00
• 88a60d71ab Fix kijai 2025-10-23 23:25:20 +03:00
• b3ad381a65 not all torch versions have this kijai 2025-10-23 21:12:45 +03:00
• 41168b1e82 Support light VAE kijai 2025-10-23 12:31:55 +03:00
• 7251c996d2 Load these LoRA keys too kijai 2025-10-22 20:43:45 +03:00
• 67fcf0ba52 Reduce needless torch.compile recompiles kijai 2025-10-22 13:29:13 +03:00
• aa610cde2b VACE: Support per context window reference images kijai 2025-10-22 11:24:30 +03:00
• 6b286552b7 MocHa: remove possibly alpha channel from inputs kijai 2025-10-21 23:26:17 +03:00
• 2ea824b9ab Mocha: Fix context window slicing kijai 2025-10-21 20:33:03 +03:00
• 001040abc7 Update nodes.py kijai 2025-10-21 20:27:37 +03:00
• 746d46e137 WanVideoScheduler: Always show split step sigma kijai 2025-10-21 20:14:54 +03:00
• 089329ef2e MoCha: revert some RoPE changes kijai 2025-10-21 20:12:40 +03:00
• f653544f7b Merge branch 'kijai:main' into custom-sigmas cmeka 2025-10-21 12:48:02 -04:00
• 425035d810 Fix custom sigmas for supported schedulers cmeka 2025-10-21 12:46:58 -04:00
• 5c1f64197e Check for actual scale_weights even if the model for some reason doesn't have the scaled_fp8 key kijai 2025-10-21 19:40:50 +03:00
• 54e938cd70 MoCha: Fix context window mask kijai 2025-10-21 19:40:01 +03:00
• 7495db7669 Update nodes_sampler.py kijai 2025-10-21 19:15:57 +03:00
• 4e36aee658 MoCha: experimental context windows support kijai 2025-10-21 18:32:40 +03:00
• 56120d633e Update nodes.py kijai 2025-10-21 17:59:37 +03:00
• 1f0861b649 MoCha: modify RoPE function to be more torch.compile friendly kijai 2025-10-21 17:54:25 +03:00
• 7916f89c33 Add updated MoCha example for lower VRAM kijai 2025-10-21 17:27:41 +03:00
• e294f417c0 Merge pull request #1501 from unrealMJ/mocha Jukka Seppänen 2025-10-21 17:17:21 +03:00
• ee011e66e9 update workflow unrealMJ 2025-10-21 17:28:50 +08:00
• d7fc563581 add mocha workflow unrealMJ 2025-10-21 10:00:51 +08:00
• 88defbfdd1 add MoCha unrealMJ 2025-10-21 09:40:54 +08:00
• 74f33df658 Merge pull request #1493 from HM-RunningHub/fix-uni3c-context-options Jukka Seppänen 2025-10-20 22:33:38 +03:00
• 487c400e8e Fix Uni3C + Context Options compatibility issue (Issue #1491) wenjian 2025-10-21 02:31:47 +08:00
• 200f6943e3 Add sageattn mode that allows torch.compile kijai 2025-10-20 15:16:43 +03:00
• 8081e1337c Fix downcasting for fp8_fast and add error to indicate you can't do scaled downcast currently kijai 2025-10-20 13:46:05 +03:00
• 9ecb80a58b Support loading VACE LoRAs (for ditto) kijai 2025-10-20 11:49:49 +03:00
• 967d15321d fix rcm scheduler check kijai 2025-10-19 19:49:24 +03:00
• 9cd79d3d4a Add experimental rCM scheduler kijai 2025-10-19 19:34:17 +03:00
• 7ecd55e92a bump version kijai 2025-10-19 18:13:09 +03:00
• 99c0efdb45 Expose force_parameter_static_shapes in torch.compile options kijai 2025-10-19 18:01:03 +03:00
• a721fe3d0f Better (hopefully) FlashVSR frame selection kijai 2025-10-19 13:49:40 +03:00
• e8ecda7240 Update wanvideo_1_3B_FlashVSR_upscale_example.json kijai 2025-10-18 16:16:25 +03:00
• 7d88316bf8 Fix for the case of applying unmerged lora to already compiled model kijai 2025-10-18 14:00:18 +03:00
• 6bcbfc54cf Fix diffusion forcing kijai 2025-10-17 18:20:45 +03:00
• 55463a6290 Fix audio_cfg_scale kijai 2025-10-17 18:12:54 +03:00
• 6aa0446706 Ovi: fix decoded output audio shape kijai 2025-10-16 18:03:42 +03:00
• 848ceb2beb Don't require whole facexlib for lynx kijai 2025-10-16 17:51:21 +03:00
• cc06d71ca0 Workaround for bug in pytorch 2.9.0 that makes the VAE use crazy amounts of VRAM kijai 2025-10-16 17:10:05 +03:00
• e221fdd7ac Fix normal tiny vae kijai 2025-10-16 14:54:00 +03:00
• c876108059 Update fp8_e4m3fn compile warning on older arch to indicate it should now work with latest Triton kijai 2025-10-16 13:08:42 +03:00