Commit Graph

  • 1db597f39c unlock max length for some text encoders (t5 and llama) power88 2025-04-20 12:21:48 +08:00
  • da3d136cb7 Use empty t5 of size 128 for hidream, seems to give closer results. comfyanonymous 2025-04-19 19:49:40 -04:00
  • fd27494441 Use empty t5 of size 128 for hidream, seems to give closer results. comfyanonymous 2025-04-19 19:49:40 -04:00
  • 6f35198397 Hidream: Allow loading hidream text encoders in CLIPLoader and DualCLIPLoader (#7676) power88 2025-04-20 07:47:30 +08:00
  • f43e1d7f41
    Hidream: Allow loading hidream text encoders in CLIPLoader and DualCLIPLoader (#7676) power88 2025-04-20 07:47:30 +08:00
  • 9e642a9b31 reformat code for ruff check. power88 2025-04-20 07:10:03 +08:00
  • 6211d2be5a Merge branch 'master' into worksplit-multigpu Jedrzej Kosinski 2025-04-19 17:36:23 -05:00
  • 8be711715c Make unload_all_models account for all devices Jedrzej Kosinski 2025-04-19 17:35:54 -05:00
  • 8ed96e73f3
    Merge a2a8d5195d0aa7139ef21961e99c5ddcf8cf81cd into 4486b0d0ff536b32100863f68e870ba18bf3d051 Silver 2025-04-19 17:25:15 -04:00
  • d892f70645 Update CODEOWNERS and add christian-byrne (#7663) Yoland Yan 2025-04-19 14:23:31 -07:00
  • 4486b0d0ff
    Update CODEOWNERS and add christian-byrne (#7663) Yoland Yan 2025-04-19 14:23:31 -07:00
  • cd77ded0c7
    Merge branch 'comfyanonymous:master' into master patientx 2025-04-19 23:20:10 +03:00
  • da12bcd594 Fix hard crash when the spiece tokenizer path is bad. comfyanonymous 2025-04-19 15:55:43 -04:00
  • 636d4bfb89 Fix hard crash when the spiece tokenizer path is bad. comfyanonymous 2025-04-19 15:55:43 -04:00
  • 39c599cace Add wanfun template workflows. (#7678) Robin Huang 2025-04-19 12:21:46 -07:00
  • dc300a4569
    Add wanfun template workflows. (#7678) Robin Huang 2025-04-19 12:21:46 -07:00
  • 7e49ba6cbc
    Add wanfun template workflows. Robin Huang 2025-04-19 12:11:11 -07:00
  • b45d782dd8 Hidream: Allow partial loading text encoders power88 2025-04-19 21:17:03 +08:00
  • 077148fe18
    Update CODEOWNERS with more files and folders yoland68-more-owner-updates Yoland Yan 2025-04-18 16:14:17 -07:00
  • 07157688d4
    Update CODEOWNERS and add christian-byrne Yoland Yan 2025-04-18 15:44:18 -07:00
  • e13cbbeeb9
    Merge branch 'comfyanonymous:master' into master patientx 2025-04-19 00:54:07 +03:00
  • 682a70bcf9
    Update zluda.py patientx 2025-04-19 00:53:59 +03:00
  • f2f7404152 Changed prints out for logging.info envy-ai 2025-04-18 16:59:45 -04:00
  • b5cccf1325 Merge branch 'master' into worksplit-multigpu Jedrzej Kosinski 2025-04-18 15:39:34 -05:00
  • 80378f5b86 KSampler: use the same noise samples across a batch Devin J. Pohly 2025-04-18 13:20:01 -05:00
  • 459b3a7fbf [BugFix] Update frontend to 1.16.9 (#7655) Chenlei Hu 2025-04-18 15:12:42 -04:00
  • f3b09b9f2d
    [BugFix] Update frontend to 1.16.9 (#7655) Chenlei Hu 2025-04-18 15:12:42 -04:00
  • a2a8d5195d Make CLIP nodes input 'type' more dynamic silveroxides 2025-04-18 19:18:36 +02:00
  • ce96a9f77f
    [BugFix] Update frontend to 1.16.9 Chenlei Hu 2025-04-18 11:48:13 -04:00
  • 159f3460ba
    Merge branch 'comfyanonymous:master' into master patientx 2025-04-18 15:40:02 +03:00
  • 5f7be8110b
    Merge branch 'comfyanonymous:master' into chroma-support Silver 2025-04-18 13:29:57 +02:00
  • a9f208ddce sync chroma-support with upstream via syncbranch patch silveroxides 2025-04-18 13:29:09 +02:00
  • 6df1402bcb Increase freq_cutoff in FreSca node. comfyanonymous 2025-04-18 03:16:16 -04:00
  • 7ecd5e9614 Increase freq_cutoff in FreSca node. desktop-release-apr222025 comfyanonymous 2025-04-18 03:16:16 -04:00
  • f9dd3567b8 Replace CLIPType if with getattr (#7589) City 2025-04-18 08:53:36 +02:00
  • 2383a39e3b
    Replace CLIPType if with getattr (#7589) City 2025-04-18 08:53:36 +02:00
  • a89710cf13 add support to output camera state (#7582) Terry Jia 2025-04-18 02:52:18 -04:00
  • 34e06bf7ec
    add support to output camera state (#7582) Terry Jia 2025-04-18 02:52:18 -04:00
  • 27c134c442 [Type] Annotate graph.get_input_info (#7386) Chenlei Hu 2025-04-17 21:02:24 -04:00
  • 55822faa05
    [Type] Annotate graph.get_input_info (#7386) Chenlei Hu 2025-04-17 21:02:24 -04:00
  • 5f477c98f7
    Update README.md patientx 2025-04-18 00:34:54 +03:00
  • 0d1aecfb43
    Merge branch 'comfyanonymous:master' into master patientx 2025-04-18 00:34:40 +03:00
  • cb8c52cbb2
    Update README.md patientx 2025-04-18 00:34:33 +03:00
  • 3d5b5b2c95
    Update pullrequest-ci-run.yml yoland68-patch-1 Yoland Yan 2025-04-17 14:17:40 -07:00
  • 319f2f81c5 Add hidream to readme. comfyanonymous 2025-04-17 16:58:27 -04:00
  • 880c205df1 Add hidream to readme. comfyanonymous 2025-04-17 16:58:27 -04:00
  • 47d2ced9f8 Make fresca work on multi dim. comfyanonymous 2025-04-17 15:46:41 -04:00
  • 3dc240d089 Make fresca work on multi dim. comfyanonymous 2025-04-17 15:46:41 -04:00
  • 2a421684c7 Add FreSca node (#7631) BVH 2025-04-18 00:54:33 +05:30
  • 19373aee75
    Add FreSca node (#7631) BVH 2025-04-18 00:54:33 +05:30
  • 0bf09baf27
    Merge branch 'comfyanonymous:master' into master patientx 2025-04-17 22:15:18 +03:00
  • 71ac9830ab
    updated comfyui-frontend version patientx 2025-04-17 22:15:11 +03:00
  • 95773a0045
    updated "comfyui-frontend version" - added "comfyui-workflow-templates" patientx 2025-04-17 22:13:36 +03:00
  • 676b934364 Run hidream with single llama text encoder envy-ai 2025-04-17 14:47:26 -04:00
  • ba51da60b4 ComfyUI version 0.3.29 comfyanonymous 2025-04-17 14:45:01 -04:00
  • 93292bc450 ComfyUI version 0.3.29 v0.3.29 comfyanonymous 2025-04-17 14:45:01 -04:00
  • ea214396c5 Add FreSca node bvhari 2025-04-17 23:58:50 +05:30
  • 019d2cb0c6 Update frontend to 1.16 (Install templates as pip package) (#7623) Christian Byrne 2025-04-18 02:25:33 +08:00
  • 05d5a75cdc
    Update frontend to 1.16 (Install templates as pip package) (#7623) Christian Byrne 2025-04-18 02:25:33 +08:00
  • fb48114cf3 bump templates version to include hidream bymyself 2025-04-17 11:15:53 -07:00
  • eaa70565b1
    Update requirements.txt Chenlei Hu 2025-04-17 13:55:03 -04:00
  • d9211bbf1e
    Merge branch 'comfyanonymous:master' into master patientx 2025-04-17 20:51:12 +03:00
  • 0d75ce7ba3 Add WanFirstLastFrameToVideo node to use the new model. comfyanonymous 2025-04-17 13:18:43 -04:00
  • eba7a25e7a Add WanFirstLastFrameToVideo node to use the new model. comfyanonymous 2025-04-17 13:18:43 -04:00
  • 5d18565fee
    Merge 4c0f5f71e8a3145038c1c4c6690499813aa27e27 into dbcfd092a29c272696bae856d943005fc0cc3036 ruucm 2025-04-18 00:59:29 +08:00
  • b16e60eafd Set default context_img_len to 257 comfyanonymous 2025-04-17 12:42:34 -04:00
  • dbcfd092a2 Set default context_img_len to 257 comfyanonymous 2025-04-17 12:42:34 -04:00
  • 9601e9e909 Support loading WAN FLF model. comfyanonymous 2025-04-17 12:04:48 -04:00
  • c14429940f Support loading WAN FLF model. comfyanonymous 2025-04-17 12:04:48 -04:00
  • 0191baf8f3 Add bit_depth option to SaveAudio for 16-bit/24-bit FLAC support MuziekMagie 2025-04-17 15:20:06 +02:00
  • 6bbe3b19d4
    Merge branch 'comfyanonymous:master' into master patientx 2025-04-17 14:51:08 +03:00
  • 89aec08623 Don't hardcode length of context_img in wan code. comfyanonymous 2025-04-17 06:25:39 -04:00
  • 0d720e4367 Don't hardcode length of context_img in wan code. comfyanonymous 2025-04-17 06:25:39 -04:00
  • 7e98854b14
    Merge branch 'comfyanonymous:master' into chroma-support Silver 2025-04-17 11:09:24 +02:00
  • 7950c510e7
    Merge branch 'comfyanonymous:master' into master patientx 2025-04-17 10:10:03 +03:00
  • 2a54a904f4 Merge branch 'master' into worksplit-multigpu Jedrzej Kosinski 2025-04-16 19:26:48 -05:00
  • 61ce29af96 install templates as pip package bymyself 2025-04-16 16:25:38 -07:00
  • 7223e38263 Make hidream work with any latent resolution. comfyanonymous 2025-04-16 18:34:14 -04:00
  • 1fc00ba4b6 Make hidream work with any latent resolution. comfyanonymous 2025-04-16 18:34:14 -04:00
  • d05e9b8298
    Merge branch 'comfyanonymous:master' into master patientx 2025-04-17 01:22:11 +03:00
  • f26cf7d5fa Limit T5 to 128 tokens for HiDream: #7620 comfyanonymous 2025-04-16 18:07:55 -04:00
  • 9899d187b1 Limit T5 to 128 tokens for HiDream: #7620 comfyanonymous 2025-04-16 18:07:55 -04:00
  • ed6f92c975 Merge branch 'master' into worksplit-multigpu Jedrzej Kosinski 2025-04-16 16:53:57 -05:00
  • 5ea8378fc6 Reuse code from flux model. comfyanonymous 2025-04-16 17:43:55 -04:00
  • f00f340a56 Reuse code from flux model. comfyanonymous 2025-04-16 17:43:55 -04:00
  • 503292b3c0
    Merge branch 'comfyanonymous:master' into master patientx 2025-04-16 23:26:10 +03:00
  • 3bcade9e24 [Type] Mark input options NotRequired (#7614) Chenlei Hu 2025-04-16 15:41:00 -04:00
  • cce1d9145e
    [Type] Mark input options NotRequired (#7614) Chenlei Hu 2025-04-16 15:41:00 -04:00
  • adc66c0698 Merge branch 'master' into worksplit-multigpu Jedrzej Kosinski 2025-04-16 14:23:56 -05:00
  • bc388f8ffa [HiDream] Hard-cap T5 embeddings to 128 length tdrussell 2025-04-16 13:42:36 -05:00
  • 7df5dbbee7
    Update README.md patientx 2025-04-16 17:08:08 +03:00
  • acaeca4344
    Merge pull request #3 from silveroxides/intermediary Silver 2025-04-16 16:00:01 +02:00
  • c791e957a6 refactor for upstream sync silveroxides 2025-04-16 15:58:37 +02:00
  • 69a6147eef merge chroma-support with upstream silveroxides 2025-04-16 15:49:24 +02:00
  • 72233ef320
    ️ Speed up function state_dict_prefix_replace by 127% Here's an optimized version of your Python function. The primary changes are to minimize the creation of intermediate lists and to use dictionary comprehensions for more efficient data manipulation. codeflash-ai[bot] 2025-04-16 10:21:07 +00:00
  • 765224ad78
    Merge branch 'comfyanonymous:master' into master patientx 2025-04-16 12:10:01 +03:00
  • 554e371cad Fix issue on old torch. comfyanonymous 2025-04-16 04:53:56 -04:00
  • b4dc03ad76 Fix issue on old torch. comfyanonymous 2025-04-16 04:53:56 -04:00
  • 03df573995 [Type] Mark input options NotRequired Chenlei Hu 2025-04-15 21:20:46 -04:00
  • 2ff056431d
    ️ Speed up method PixelNorm.forward by 7% To optimize the runtime of this program, we can leverage some of PyTorch's functions for better performance. Specifically, we can use torch.rsqrt and torch.mean wisely to optimize the normalization calculation. This can be beneficial from a performance perspective since certain operations might be optimized internally. codeflash-ai[bot] 2025-04-15 22:19:03 +00:00