Commits

Breadcrumbs

History for

onmain

User selector

All users

All time

Commit History

Commits on Feb 13, 2026

• Sunset Python 3.8 & get rid of explicit `typing` exports where possible (#12524)

  Show description for 2843b3d
  

  sayakpaul

  and

  DN6

  authored
  2843b3d

  Copy full SHA for 2843b3d

Commits on Jan 16, 2026

• [core] gracefully error out when attn-backend x cp combo isn't supported. (#12832)

  Show description for ebf891a
  

  sayakpaul

  and

  DN6

  authored
  ebf891a

  Copy full SHA for ebf891a

Commits on Jan 14, 2026

• `disable_mmap` in pipeline `from_pretrained` (#12854)

  Show description for 1ecfbfe
  

  hlky

  and

  DN6

  authored
  1ecfbfe

  Copy full SHA for 1ecfbfe

• Remove 8bit device restriction (#12972)

  Show description for d7fa445
  

  3 peopleauthored
  d7fa445

  Copy full SHA for d7fa445

Commits on Jan 9, 2026

• [core] Handle progress bar and logging in distributed environments (#12806)

  Show description for 441b69e
  sayakpaul

  authored
  441b69e

  Copy full SHA for 441b69e

Commits on Dec 5, 2025

• Fix broken group offloading with block_level for models with standalone layers (#12692)

  Show description for f12d161
  

  3 peopleauthored
  f12d161

  Copy full SHA for f12d161

Commits on Nov 19, 2025

• [core] Refactor hub attn kernels (#12475)

  Show description for ab71f3c
  

  3 peopleauthored
  ab71f3c

  Copy full SHA for ab71f3c

Commits on Nov 10, 2025

• Fix Context Parallel validation checks (#12446)

  Show description for 8d45f21
  

  DN6

  and

  sayakpaul

  authored
  8d45f21

  Copy full SHA for 8d45f21

Commits on Oct 22, 2025

• Fix: Add _skip_keys for AutoencoderKLWan (#12523)

  Show description for bec2d8e
  yiyixuxu

  authored
  bec2d8e

  Copy full SHA for bec2d8e

Commits on Oct 17, 2025

• docs: cleanup of runway model (#12503)

  Show description for 1b456bd
  EazyAl

  authored
  1b456bd

  Copy full SHA for 1b456bd

Commits on Sep 30, 2025

• [docs] Migrate syntax (#12390)

  Show description for cc5b31f
  stevhliu

  authored
  cc5b31f

  Copy full SHA for cc5b31f

Commits on Sep 24, 2025

• Context Parallel w/ Ring & Ulysses & Unified Attention (#11941)

  Show description for dcb6dd9
  

  3 peopleauthored
  dcb6dd9

  Copy full SHA for dcb6dd9

Commits on Aug 14, 2025

• support `hf_quantizer` in cache warmup. (#12043)

  Show description for 58bf268
  sayakpaul

  authored
  58bf268

  Copy full SHA for 58bf268

• make parallel loading flag a part of constants. (#12137)

  sayakpaul

  authored
  123506e

  Copy full SHA for 123506e

Commits on Aug 13, 2025

• [core] parallel loading of shards (#12028)

  Show description for baa9b58
  

  sayakpaul

  and

  DN6

  authored
  baa9b58

  Copy full SHA for baa9b58

Commits on Jul 29, 2025

• Fix huggingface-hub failing tests (#11994)

  Show description for edcbe80
  asomoza

  authored
  edcbe80

  Copy full SHA for edcbe80

Commits on Jul 22, 2025

• Remove logger warnings for attention backends and hard error during runtime instead (#11967)

  Show description for e46e139
  a-r-r-o-w

  authored
  e46e139

  Copy full SHA for e46e139

Commits on Jul 17, 2025

• [refactor] Flux/Chroma single file implementation + Attention Dispatcher (#11916)

  Show description for 18c8f10
  

  a-r-r-o-w

  and

  DN6

  authored
  18c8f10

  Copy full SHA for 18c8f10

Commits on Jul 16, 2025

• Add SkyReels V2: Infinite-Length Film Generative Model (#11518)

  Show description for 7298bdd
  

  3 peopleauthored
  7298bdd

  Copy full SHA for 7298bdd

• Fixed bug: Uncontrolled recursive calls that caused an infinite loop when loading certain pipelines containing Transformer2DModel (#11923)

  Show description for c5d6e0b
  

  lengmo1996

  and

  asomoza

  authored
  c5d6e0b

  Copy full SHA for c5d6e0b

Commits on Jul 15, 2025

• Remove device synchronization when loading weights (#11927)

  Show description for b73c738
  a-r-r-o-w

  authored
  b73c738

  Copy full SHA for b73c738

Commits on Jul 11, 2025

• Speedup model loading by 4-5x ⚡ (#11904)

  Show description for c903527
  

  a-r-r-o-w

  and

  SunMarc

  authored
  c903527

  Copy full SHA for c903527

Commits on Jun 26, 2025

• [rfc][compile] compile method for DiffusionPipeline (#11705)

  Show description for d93381c
  

  3 peopleauthored
  d93381c

  Copy full SHA for d93381c

Commits on Jun 19, 2025

• make group offloading work with disk/nvme transfers (#11682)

  Show description for 85a916b
  sayakpaul

  authored
  85a916b

  Copy full SHA for 85a916b

Commits on Jun 11, 2025

• [tests] model-level `device_map` clarifications (#11681)

  Show description for 9154566
  sayakpaul

  authored
  9154566

  Copy full SHA for 9154566

Commits on May 13, 2025

• fix: remove `torch_dtype="auto"` option from docstrings (#11513)

  Show description for f8d4a1e
  

  johannaSommer

  and

  DN6

  authored
  f8d4a1e

  Copy full SHA for f8d4a1e

Commits on Apr 8, 2025

• [feat] implement `record_stream` when using CUDA streams during group offloading (#11081)

  Show description for 4b27c4a
  

  sayakpaul

  and

  a-r-r-o-w

  authored
  4b27c4a

  Copy full SHA for 4b27c4a

Commits on Apr 2, 2025

• Revert `save_model` in ModelMixin save_pretrained and use safe_serialization=False in test (#11196)

  hlky

  authored
  da857be

  Copy full SHA for da857be

• allow models to run with a user-provided dtype map instead of a single dtype (#10301)

  Show description for d8c617c
  

  hlky

  and

  sayakpaul

  authored
  d8c617c

  Copy full SHA for d8c617c

Commits on Mar 21, 2025

• Don't override `torch_dtype` and don't use when `quantization_config` is set (#11039)

  Show description for a7d53a5
  

  hlky

  and

  sayakpaul

  authored
  a7d53a5

  Copy full SHA for a7d53a5

• [core] FasterCache (#10163)

  Show description for 844221a
  

  a-r-r-o-w

  and

  github-actions[bot]

  authored
  844221a

  Copy full SHA for 844221a

Commits on Mar 20, 2025

• Provide option to reduce CPU RAM usage in Group Offload (#11106)

  Show description for 2c1ed50
  DN6

  authored
  2c1ed50

  Copy full SHA for 2c1ed50

Commits on Mar 2, 2025

• [Alibaba Wan Team] continue on #10921 Wan2.1 (#10922)

  Show description for 2d8a41c
  

  4 peopleauthored
  2d8a41c

  Copy full SHA for 2d8a41c

Commits on Feb 24, 2025

• Fix `torch_dtype` in Kolors text encoder with `transformers` v4.49 (#10816)

  Show description for 6f74ef5
  hlky

  authored
  6f74ef5

  Copy full SHA for 6f74ef5

Commits on Feb 20, 2025

• Remove print statements (#10836)

  Show description for f8b54cf
  a-r-r-o-w

  authored
  f8b54cf

  Copy full SHA for f8b54cf