andreas/ComfyUI
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Commit Graph

  • d49420b3c7 LongCat-Image edit (#13003) Talmaj 2026-03-22 04:51:05 +01:00
  • ebf6b52e32 ComfyUI v0.18.1 comfyanonymous 2026-03-21 22:32:16 -04:00
  • 25b6d1d629 wan: vae: Fix light/color change (#13101) rattus 2026-03-21 15:44:35 -07:00
  • 11c15d8832 Fix fp16 intermediates giving different results. (#13100) comfyanonymous 2026-03-21 14:53:25 -07:00
  • b5d32e6ad2 Fix sampling issue with fp16 intermediates. (#13099) comfyanonymous 2026-03-21 14:47:42 -07:00
  • a11f68dd3b Fix canny node not working with fp16. (#13085) comfyanonymous 2026-03-20 20:15:50 -07:00
  • dc719cde9c ComfyUI version 0.18.0 comfyanonymous 2026-03-20 20:09:15 -04:00
  • 87cda1fc25 Move inline comfy.context_windows imports to top-level in model_base.py (#13083) Jedrzej Kosinski 2026-03-20 17:03:42 -07:00
  • 45d5c83a30 Make EmptyImage node follow intermediate device/dtype. (#13079) comfyanonymous 2026-03-20 13:08:26 -07:00
  • c646d211be feat(api-nodes): add Quiver SVG nodes (#13047) Alexander Piskun 2026-03-20 21:23:16 +02:00
  • 589228e671 Add slice_cond and per-model context window cond resizing (#12645) drozbay 2026-03-19 21:42:42 -06:00
  • e4455fd43a [API Nodes] mark seedream-3-0-t2i and seedance-1-0-lite models as deprecated (#13060) Alexander Piskun 2026-03-20 05:05:01 +02:00
  • f49856af57 ltx: vae: Fix missing init variable (#13074) rattus 2026-03-19 19:34:58 -07:00
  • 82b868a45a Fix VRAM leak in tiler fallback in video VAEs (#13073) rattus 2026-03-19 19:30:27 -07:00
  • 8458ae2686 Revert "fix: run text encoders on MPS GPU instead of CPU for Apple Silicon (#…" (#13070) comfyanonymous 2026-03-19 12:27:55 -07:00
  • fd0261d2bc Reduce tiled decode peak memory (#13050) Jukka Seppänen 2026-03-19 19:29:34 +02:00
  • ab14541ef7 memory: Add more exclusion criteria to pinned read (#13067) rattus 2026-03-19 10:03:20 -07:00
  • 6589562ae3 ltx: vae: implement chunked encoder + CPU IO chunking (Big VRAM reductions) (#13062) rattus 2026-03-19 10:01:12 -07:00
  • fabed694a2 ltx: vae: implement chunked encoder + CPU IO chunking (Big VRAM reductions) (#13062) rattus 2026-03-19 09:58:47 -07:00
  • f6b869d7d3 fp16 intermediates doen't work for some text enc models. (#13056) comfyanonymous 2026-03-18 16:42:28 -07:00
  • 56ff88f951 Fix regression. (#13053) comfyanonymous 2026-03-18 15:35:25 -07:00
  • 9fff091f35 Further Reduce LTX VAE decode peak RAM usage (#13052) Jukka Seppänen 2026-03-19 00:32:26 +02:00
  • dcd659590f Make more intermediate values follow the intermediate dtype. (#13051) comfyanonymous 2026-03-18 15:14:18 -07:00
  • b67ed2a45f Update comfyui-frontend-package version to 1.41.21 (#13035) Alexander Brown 2026-03-18 13:36:39 -07:00
  • 06957022d4 fix(api-nodes): add support for "thought_image" in Nano Banana 2 and corrected price badges (#13038) Alexander Piskun 2026-03-18 19:21:58 +02:00
  • b941913f1d fix: run text encoders on MPS GPU instead of CPU for Apple Silicon (#12809) Anton Bukov 2026-03-18 05:21:32 +04:00
  • cad24ce262 cascade: remove dead weight init code (#13026) rattus 2026-03-17 17:59:10 -07:00
  • 68d542cc06 Fix case where pixel space VAE could cause issues. (#13030) comfyanonymous 2026-03-17 17:46:22 -07:00
  • 735a0465e5 Inplace VAE output processing to reduce peak RAM consumption. (#13028) Jukka Seppänen 2026-03-18 02:20:49 +02:00
  • 8b9d039f26 bump manager version to 4.1b6 (#13022) Dr.Lt.Data 2026-03-18 07:17:03 +09:00
  • 035414ede4 Reduce WAN VAE VRAM, Save use cases for OOM/Tiler (#13014) rattus 2026-03-17 14:34:39 -07:00
  • 1a157e1f97 Reduce LTX VAE VRAM usage and save use cases from OOMs/Tiler (#13013) rattus 2026-03-17 14:32:43 -07:00
  • ed7c2c6579 Mark weight_dtype as advanced input in Load Diffusion Model node (#12769) Christian Byrne 2026-03-17 07:24:00 -07:00
  • 379fbd1a82 chore: update workflow templates to v0.9.26 (#13012) ComfyUI Wiki 2026-03-17 12:53:18 +08:00
  • 8cc746a864 fix: disable SageAttention for Hunyuan3D v2.1 DiT (#12772) Paulo Muggler Moreira 2026-03-17 03:27:27 +01:00
  • 9a870b5102 fix: atomic writes for userdata to prevent data loss on crash (#12987) Christian Byrne 2026-03-16 18:56:35 -07:00
  • ca17fc8355 Fix potential issue. (#13009) comfyanonymous 2026-03-16 18:38:40 -07:00
  • 20561aa919 [Trainer] FP4, 8, 16 training by native dtype support and quant linear autograd function (#12681) Kohaku-Blueleaf 2026-03-17 09:31:50 +08:00
  • 7a16e8aa4e Add --enable-dynamic-vram options to force enable it. (#13002) comfyanonymous 2026-03-16 13:50:13 -07:00
  • b202f842af Skip running model finalizers at exit (#12994) blepping 2026-03-16 14:00:42 -06:00
  • 7d5f5252c3 ci: add check to block AI agent Co-authored-by trailers in PRs (#12799) Christian Byrne 2026-03-16 12:53:13 -07:00
  • 2bd4d82b4f feat(assets): align local API with cloud spec (#12863) Luke Mino-Altherr 2026-03-16 15:34:04 -04:00
  • 593be209a4 feat: add essentials_category to nodes and blueprints for Essentials tab (#12573) Christian Byrne 2026-03-15 16:18:04 -07:00
  • 3814bf4454 Enable Pytorch Attention for gfx1150 (#12973) lostdisc 2026-03-15 15:45:30 -04:00
  • d062becb33 Make EmptyLatentImage follow intermediate dtype. (#12974) comfyanonymous 2026-03-15 12:37:27 -07:00
  • e84a200a3c ops: opt out of deferred weight init if subclassed (#12967) rattus 2026-03-15 11:49:49 -07:00
  • 192cb8eeb9 bump manager version to 4.1b5 (#12957) Dr.Lt.Data 2026-03-16 03:48:56 +09:00
  • 0904cc3fe5 LTXV: Accumulate VAE decode results on intermediate_device (#12955) Jukka Seppänen 2026-03-15 03:09:09 +02:00
  • 4941cd046e Update comfyui-frontend-package to version 1.41.20 (#12954) comfyanonymous 2026-03-14 16:53:31 -07:00
  • c711b8f437 Add --fp16-intermediates to use fp16 for intermediate values between nodes (#12953) comfyanonymous 2026-03-14 16:18:19 -07:00
  • 1c5db7397d feat: Support mxfp8 (#12907) Jukka Seppänen 2026-03-15 00:36:29 +02:00
  • e0982a7174 fix: use no-store cache headers to prevent stale frontend chunks (#12911) Christian Byrne 2026-03-14 15:25:09 -07:00
  • 4c4be1bba5 comfy-aimdo 0.2.12 (#12941) rattus 2026-03-14 07:53:00 -07:00
  • 16cd8d8a8f Update README. (#12931) comfyanonymous 2026-03-13 19:33:28 -07:00
  • 7810f49702 comfy aimdo 0.2.11 + Improved RAM Pressure release strategies - Windows speedups (#12925) rattus 2026-03-13 19:18:08 -07:00
  • e1f10ca093 bump manager version to 4.1b4 (#12930) Dr.Lt.Data 2026-03-14 09:14:27 +09:00
  • 6cd35a0c5f Bump comfyui-frontend-package to 1.41.19 (#12923) Comfy Org PR Bot 2026-03-14 03:31:25 +09:00
  • f9ceed9eef fix(api-nodes): Tencent TextToModel and ImageToModel nodes (#12680) Alexander Piskun 2026-03-13 19:10:40 +02:00
  • 4a8cf359fe Revert "Revert "feat: Add CacheProvider API for external distributed caching"" (#12915) Deep Mehta 2026-03-12 21:17:50 -07:00
  • 63d1bbdb40 ComfyUI v0.17.0 comfyanonymous 2026-03-12 20:41:48 -04:00
  • 5df1427124 Fix audio extraction and truncation bugs (#12652) PxTicks 2026-03-13 00:44:15 +00:00
  • d1d53c14be Revert "feat: Add CacheProvider API for external distributed caching (#12056)" (#12912) comfyanonymous 2026-03-12 17:21:23 -07:00
  • af7b4a921d feat: Add CacheProvider API for external distributed caching (#12056) Deep Mehta 2026-03-12 16:09:07 -07:00
  • 8d9faaa181 Update requirements.txt (#12910) Christian Byrne 2026-03-12 15:14:59 -07:00
  • 47e1e316c5 Lower kv cache memory usage. (#12909) comfyanonymous 2026-03-12 13:54:38 -07:00
  • 712411d539 chore: update workflow templates to v0.9.21 (#12908) ComfyUI Wiki 2026-03-13 03:16:54 +08:00
  • 3fa8c5686d fix: use frontend-compatible format for Float gradient_stops (#12789) Terry Jia 2026-03-12 10:14:28 -07:00
  • 73d9599495 add painter node (#12294) Terry Jia 2026-03-12 09:55:29 -07:00
  • 44f1246c89 Support flux 2 klein kv cache model: Use the FluxKVCache node. (#12905) comfyanonymous 2026-03-12 08:30:50 -07:00
  • 8f9ea49571 Bump comfy-kitchen version to 0.2.8 (#12895) comfyanonymous 2026-03-11 21:17:31 -07:00
  • 9ce4c3dd87 Bump comfyui-frontend-package to 1.41.16 (#12894) Comfy Org PR Bot 2026-03-12 10:16:30 +09:00
  • abc87d3669 Bump comfyui-frontend-package to 1.41.15 (#12891) Comfy Org PR Bot 2026-03-12 06:04:51 +09:00
  • f6274c06b4 Fix issue with batch_size > 1 on some models. (#12892) comfyanonymous 2026-03-11 13:37:31 -07:00
  • 4f4f8659c2 fix: guard torch.AcceleratorError for compatibility with torch < 2.8.0 (#12874) Adi Borochov 2026-03-11 19:04:13 +02:00
  • 3365008dfe feat(api-nodes): add Reve Image nodes (#12848) Alexander Piskun 2026-03-11 18:53:55 +02:00
  • 980621da83 comfy-aimdo 0.2.10 (#12890) rattus 2026-03-11 08:49:38 -07:00
  • 9642e4407b Add pre attention and post input patches to qwen image model. (#12879) comfyanonymous 2026-03-10 21:09:35 -07:00
  • 3ad36d6be6 Allow model patches to have a cleanup function. (#12878) comfyanonymous 2026-03-10 17:09:12 -07:00
  • 8086468d2a main: switch on faulthandler (#12868) rattus 2026-03-10 09:05:31 -07:00
  • 535c16ce6e Widen OOM_EXCEPTION to AcceleratorError form (#12835) rattus 2026-03-09 21:41:02 -07:00
  • a912809c25 model_detection: deep clone pre edited edited weights (#12862) rattus 2026-03-09 20:50:10 -07:00
  • c4fb0271cd Add a way for nodes to add pre attn patches to flux model. (#12861) comfyanonymous 2026-03-09 20:37:58 -07:00
  • 740d998c9c fix(manager): improve install guidance when comfyui-manager is not installed (#12810) Dr.Lt.Data 2026-03-10 11:49:31 +09:00
  • 814dab9f46 Update workflow templates to v0.9.18 (#12857) ComfyUI Wiki 2026-03-10 10:03:22 +08:00
  • 06f85e2c79 Fix text encoder lora loading for wrapped models (#12852) Jukka Seppänen 2026-03-09 22:08:51 +02:00
  • e4b0bb8305 Import assets seeder later, print some package versions. (#12841) comfyanonymous 2026-03-08 13:25:30 -07:00
  • 7723f20bbe comfy-aimdo 0.2.9 (#12840) rattus 2026-03-08 13:17:40 -07:00
  • 29b24cb517 refactor(assets): modular architecture + async two-phase scanner & background seeder (#12621) Luke Mino-Altherr 2026-03-07 17:37:25 -08:00
  • a7a6335be5 ComfyUI v0.16.4 comfyanonymous 2026-03-07 16:52:39 -05:00
  • bcf1a1fab1 mm: reset_cast_buffers: sync compute stream before free (#12822) rattus 2026-03-07 09:38:08 -08:00
  • 6ac8152fc8 chore: update workflow templates to v0.9.11 (#12821) ComfyUI Wiki 2026-03-07 15:54:09 +08:00
  • afc00f0055 Fix requirements version. (#12817) comfyanonymous 2026-03-06 17:10:53 -08:00
  • d69d30819b Don't run TE on cpu when dynamic vram enabled. (#12815) comfyanonymous 2026-03-06 16:11:16 -08:00
  • f466b06601 Fix fp16 audio encoder models (#12811) rattus 2026-03-06 15:20:07 -08:00
  • 34e55f0061 feat(api-nodes): add Gemini 3.1 Flash Lite model to LLM node (#12803) Alexander Piskun 2026-03-06 19:54:27 +02:00
  • 3b93d5d571 feat(api-nodes): add TencentSmartTopology node (#12741) Alexander Piskun 2026-03-06 11:04:48 +02:00
  • e544c65db9 feat: add Math Expression node with simpleeval evaluation (#12687) Dante 2026-03-06 11:51:28 +09:00
  • 1c21828236 ComfyUI v0.16.3 comfyanonymous 2026-03-05 17:25:49 -05:00
  • 58017e8726 feat: add causal_fix parameter to add_keyframe_index and append_keyframe (#12797) Tavi Halperin 2026-03-05 23:51:20 +02:00
  • 17b43c2b87 LTX audio vae novram fixes. (#12796) comfyanonymous 2026-03-05 13:31:28 -08:00