Commit Graph

  • dea899f221
    Unload weights if vram usage goes up between runs. (#10690) comfyanonymous 2025-11-09 15:51:33 -08:00
  • e632e5de28
    Add logging for model unloading. (#10692) comfyanonymous 2025-11-09 15:06:39 -08:00
  • e920a84bd7
    Add logging for model unloading. comfyanonymous 2025-11-09 17:59:48 -05:00
  • bff719976e Do not pin memory on Intel devices a-One-Fan 2025-11-09 23:15:06 +02:00
  • 0b58299ab0 Unload weights if vram usage goes up between runs. comfyanonymous 2025-11-09 16:02:32 -05:00
  • e07a32c9b8
    Merge branch 'master' into offloader-maifee Maifee Ul Asad 2025-11-09 17:25:46 +06:00
  • d56cb56059
    Merge branch 'comfyanonymous:master' into master patientx 2025-11-09 13:41:51 +03:00
  • 2abd2b5c20
    Make ScaleROPE node work on Flux. (#10686) comfyanonymous 2025-11-08 12:52:02 -08:00
  • 2abd6767bc Make ScaleROPE node work on Flux. comfyanonymous 2025-11-08 15:42:03 -05:00
  • 44346c4251 removed all errors Yousef Rafat 2025-11-08 19:49:02 +02:00
  • de408d5f2c Add env TORCH_AMD_CUDNN_ENABLED (fix typo) Christopher Anderson 2025-11-08 20:45:32 +11:00
  • ca73329fcd mm: allow unload of the current model Rattus 2025-11-08 16:03:39 +10:00
  • b85ed81792 add ui pbar Kohaku-Blueleaf 2025-11-08 15:42:18 +08:00
  • b3784a7da1 Rewrite training system with new io schema Kohaku-Blueleaf 2025-11-08 15:06:48 +08:00
  • 28f22a517a Rewrite dataset system with new io schema Kohaku-Blueleaf 2025-11-08 15:06:36 +08:00
  • 90ba86a95f Add env TORCH_AMD_CUDNN_ENABLED (import os) Christopher Anderson 2025-11-08 16:12:45 +11:00
  • 8e02689534
    Merge branch 'comfyanonymous:master' into master patientx 2025-11-07 20:30:21 +03:00
  • 417dbcb32f Add env TORCH_AMD_CUDNN_ENABLED Christopher Anderson 2025-11-08 03:37:46 +11:00
  • a1a70362ca
    Only unpin tensor if it was pinned by ComfyUI (#10677) comfyanonymous 2025-11-07 08:15:05 -08:00
  • 8b4abb17b2
    Only unpin tensor if it was pinned by ComfyUI comfyanonymous 2025-11-07 10:58:56 -05:00
  • d29dbbd829
    Merge branch 'comfyanonymous:master' into master patientx 2025-11-07 14:27:13 +03:00
  • a603035a4c
    Merge branch 'comfyanonymous:master' into feature/custom_nodes-envvar Sas van Gulik 2025-11-07 06:18:07 +01:00
  • cf97b033ee
    mm: guard against double pin and unpin explicitly (#10672) rattus 2025-11-07 12:20:48 +10:00
  • 1520068058 mm: guard against double pin and unpin explicitly Rattus 2025-11-07 11:23:26 +10:00
  • eb1c42f649
    Tell users they need to upload their logs in bug reports. (#10671) comfyanonymous 2025-11-06 17:24:28 -08:00
  • a1366f102b
    Tell users they need to upload their logs in bug reports. comfyanonymous 2025-11-06 20:22:44 -05:00
  • 894dcf09c2
    convert nodes_mask.py to V3 schema bigcat88 2025-11-06 13:07:17 +02:00
  • 9d1f7c5eee
    Update README.md patientx 2025-11-06 15:35:55 +03:00
  • 3ab45ae725
    Merge branch 'comfyanonymous:master' into master patientx 2025-11-06 15:35:41 +03:00
  • db70e60d01
    Update README.md patientx 2025-11-06 15:35:33 +03:00
  • 13dbabe76c
    Merge branch 'comfyanonymous:master' into feature/custom_nodes-envvar Sas van Gulik 2025-11-06 10:25:45 +01:00
  • 41b875c0ed
    Update documentation for custom nodes directory Sas van Gulik 2025-11-06 10:20:20 +01:00
  • 60163d5643
    Remove environment variable. Sas van Gulik 2025-11-06 10:12:28 +01:00
  • e05c907126
    Clarify release cycle. (#10667) comfyanonymous 2025-11-06 01:11:30 -08:00
  • 20a1b10c25
    Clarify release cycle. comfyanonymous 2025-11-06 04:10:49 -05:00
  • 8efc4303af
    Merge branch 'comfyanonymous:master' into feature/custom_nodes-envvar Sas van Gulik 2025-11-06 10:10:32 +01:00
  • fbb69af758 [add] tripo3.0 liangd 2025-11-06 14:02:23 +08:00
  • 08c28cca03
    Merge 7e5db0866ff39f06193da77bc7aa0814db89064b into 09dc24c8a982776abd5cb2f71e3d041139e1d5b2 David 2025-11-06 03:04:50 +00:00
  • 7e5db0866f
    Update cldm.py David 2025-11-05 22:04:36 -05:00
  • 09dc24c8a9
    Pinned mem also seems to work on AMD. (#10658) comfyanonymous 2025-11-05 16:11:15 -08:00
  • 6c8a4fbcd1
    Pinned mem also seems to work on AMD. comfyanonymous 2025-11-05 19:06:41 -05:00
  • 1d69245981
    Enable pinned memory by default on Nvidia. (#10656) comfyanonymous 2025-11-05 15:08:13 -08:00
  • 97f198e421
    Fix qwen controlnet regression. (#10657) comfyanonymous 2025-11-05 15:07:35 -08:00
  • 62831c5fd5
    Fix qwen controlnet regression. comfyanonymous 2025-11-05 17:57:35 -05:00
  • 884cd9f0bc Enable pinned memory by default on Nvidia. comfyanonymous 2025-11-05 17:49:51 -05:00
  • 5056a1f4d4 important fixes Yousef Rafat 2025-11-06 00:24:49 +02:00
  • c25f5d3152
    Merge branch 'comfyanonymous:master' into master patientx 2025-11-05 20:08:21 +03:00
  • 5cb77fbb18 Merge branch 'master' into dr-support-pip-cm Dr.Lt.Data 2025-11-06 00:55:10 +09:00
  • fb099a40b2 Handle ref_attn_mask with separate patch to avoid having to always return q and k from self_attn kijai 2025-11-05 17:24:13 +02:00
  • f051129c1d
    Merge branch 'comfyanonymous:master' into feature/custom_nodes-envvar Sas van Gulik 2025-11-05 11:51:28 +01:00
  • 4a010caa0d change delimiter type for seperate folders per platform. Sasbom 2025-11-05 11:26:11 +01:00
  • 764d6c43f3 allow for PATH style ';' delimited custom_node directories. Sasbom 2025-11-05 11:16:59 +01:00
  • bda0eb2448
    feat(API-nodes): move Rodin3D nodes to new client; removed old api client.py (#10645) Alexander Piskun 2025-11-05 12:16:00 +02:00
  • 84faf45f09
    Merge branch 'comfyanonymous:master' into master patientx 2025-11-05 13:07:02 +03:00
  • 7237c36fba Merge remote-tracking branch 'upstream/master' into multitalk kijai 2025-11-05 11:36:07 +02:00
  • 90b6ed7d09
    bugfix: fix typo in apply_directory for custom_nodes_directory Sas van Gulik 2025-11-05 09:29:42 +01:00
  • 175c22dad4 Merge remote-tracking branch 'origin/sfink-amd-cudnn-envvar' into sfink-amd-cudnn-envvar Christopher Anderson 2025-11-05 18:35:17 +11:00
  • 58db8864a6 Add env TORCH_AMD_CUDNN_ENABLED Christopher Anderson 2025-11-05 18:35:00 +11:00
  • f9618b7e4b Add QOL feature for changing the custom nodes folder location through cli args or a COMFYUI_CUSTOM_NODES_DIR environment variable. Sasbom 2025-11-05 07:40:32 +01:00
  • 242037fa32 Fix EliGen RoPE shape to match apply_rope1() requirements nolan4 2025-11-04 21:55:02 -08:00
  • 9792606847 added attention_mask to QwenImageTransformerBlock.forward() nolan4 2025-11-04 19:55:24 -08:00
  • 581f8fe930 Also add MatchType check to input_type in validation - will likely trigger when connecting to non-lazy stuff Jedrzej Kosinski 2025-11-04 19:54:28 -08:00
  • c4a6b389de
    Lower ltxv mem usage to what it was before previous pr. (#10643) comfyanonymous 2025-11-04 19:47:35 -08:00
  • bd78daa9a7 Make match type receive_type pass validation Jedrzej Kosinski 2025-11-04 19:41:14 -08:00
  • 62c1e888cd Lower ltxv mem usage to what it was before previous pr. comfyanonymous 2025-11-04 22:40:50 -05:00
  • 53c4aab268 Merge branch 'combo-output-fix' into v3-match-type Jedrzej Kosinski 2025-11-04 19:37:44 -08:00
  • 4cd881866b
    Use single apply_rope function across models (#10547) contentis 2025-11-05 02:10:11 +01:00
  • 7b77a0d305 Add workaround in validation.py for V3 Combo outputs not working as Combo inputs combo-output-fix Jedrzej Kosinski 2025-11-04 16:58:34 -08:00
  • 265adad858 ComfyUI version v0.3.68 v0.3.68 comfyanonymous 2025-11-04 19:42:23 -05:00
  • e31f7a1128 Fixed providing list of allowed_types Jedrzej Kosinski 2025-11-04 16:06:26 -08:00
  • 16adfe2153 restored whitespace and fixed logging nolan4 2025-11-04 15:46:52 -08:00
  • 32bd55779a Merge branch 'master' into dr-support-pip-cm Dr.Lt.Data 2025-11-05 07:42:29 +09:00
  • 7f3e4d486c
    Limit amount of pinned memory on windows to prevent issues. (#10638) comfyanonymous 2025-11-04 14:37:50 -08:00
  • 4057540efe Added output_matchtypes to generated json for v3, initial backend support for MatchType, created nodes_logic.py and added SwitchNode Jedrzej Kosinski 2025-11-04 14:30:03 -08:00
  • 704d137c90 Limit amount of pinned memory on windows to prevent issues. comfyanonymous 2025-11-04 17:26:29 -05:00
  • a389ee01bb
    caching: Handle None outputs tuple case (#10637) rattus 2025-11-05 08:14:10 +10:00
  • 78aee2df53 caching: Handle None outputs tuple case Rattus 2025-11-05 08:01:39 +10:00
  • 9e9c536c8e fixes from testing Yousef Rafat 2025-11-04 23:55:16 +02:00
  • 9c71a66790
    chore: update workflow templates to v0.2.11 (#10634) ComfyUI Wiki 2025-11-05 02:51:53 +08:00
  • 6f6781fa4b chore: update workflow templates to v0.2.11 ComfyUI Wiki 2025-11-05 02:41:02 +08:00
  • fe28a51589 chore: update workflow templates to v0.2.10 ComfyUI Wiki 2025-11-05 02:14:21 +08:00
  • 671a769dc6 Merge branch 'master' into dr-support-pip-cm Dr.Lt.Data 2025-11-04 23:25:51 +09:00
  • 11083ab58c
    Merge branch 'comfyanonymous:master' into master patientx 2025-11-04 13:09:30 +03:00
  • 88746bf3aa Merge branch 'master' into dataset-process-nodes Kohaku-Blueleaf 2025-11-04 17:25:41 +08:00
  • 650b7b0302 move all dataset related implementation to nodes_dataset Kohaku-Blueleaf 2025-11-04 12:51:40 +08:00
  • 992aa2dd8f allow trainer node to get different resolution dataset Kohaku-Blueleaf 2025-11-04 12:50:25 +08:00
  • af4b7b5edb
    More fp8 torch.compile regressions fixed. (#10625) comfyanonymous 2025-11-03 19:14:20 -08:00
  • bc3daf25f1 More fp8 torch.compile regressions fixed. comfyanonymous 2025-11-03 22:07:05 -05:00
  • 9db1425b70 fix: update folder path logic to use correct module path limon-daohaosisi 2025-11-04 10:58:07 +08:00
  • 0f4ef3afa0
    This seems to slow things down slightly on Linux. (#10624) comfyanonymous 2025-11-03 18:47:14 -08:00
  • 499f6d8a82
    This seems to slow things down slightly on Linux. comfyanonymous 2025-11-03 21:39:15 -05:00
  • 6b88478f9f
    Bring back fp8 torch compile performance to what it should be. (#10622) comfyanonymous 2025-11-03 16:22:10 -08:00
  • 34c79a1d38
    Fix recursion. comfyanonymous 2025-11-03 19:15:35 -05:00
  • 8230919053
    Remove debug print for tensor shapes comfyanonymous 2025-11-03 19:00:23 -05:00
  • 6c84d84a48 Bring back fp8 torch compile performance to what it should be. comfyanonymous 2025-11-03 18:58:19 -05:00
  • e199c8cc67
    Fixes (#10621) comfyanonymous 2025-11-03 14:58:24 -08:00
  • 3abd66aace
    Fixes comfyanonymous 2025-11-03 17:51:53 -05:00
  • 0652cb8e2d
    Speed up torch.compile (#10620) comfyanonymous 2025-11-03 14:37:12 -08:00
  • 197d778fdf
    Speed up torch.compile comfyanonymous 2025-11-03 17:33:34 -05:00
  • 958a17199a
    People should update their pytorch versions. (#10618) comfyanonymous 2025-11-03 14:08:30 -08:00