Commits

Breadcrumbs

History for

onmain

User selector

All users

All time

Commit History

Commits on Feb 18, 2025

• Fix max_shift value in flux and related functions to 1.15 (issue #10675) (#10807)

  Show description for b75b204
  puhuk

  authored
  b75b204

  Copy full SHA for b75b204

Commits on Feb 17, 2025

• [LoRA] improve lora support for flux. (#10810)

  Show description for c14057c
  sayakpaul

  authored
  c14057c

  Copy full SHA for c14057c

Commits on Feb 16, 2025

• Extend Support for callback_on_step_end for AuraFlow and LuminaText2Img Pipelines (#10746)

  Show description for 3e99b56
  

  ParagEkbote

  and

  hlky

  authored
  3e99b56

  Copy full SHA for 3e99b56

Commits on Feb 15, 2025

• CogView4 (supports different length c and uc) (#10649)

  Show description for d90cd36
  

  4 peopleauthored
  d90cd36

  Copy full SHA for d90cd36

• follow-up refactor on lumina2 (#10776)

  Show description for 69f919d
  yiyixuxu

  authored
  69f919d

  Copy full SHA for 69f919d

Commits on Feb 14, 2025

• [FIX] check_inputs function in lumina2 (#10784)

  SahilCarterr

  authored
  a6b843a

  Copy full SHA for a6b843a

• Module Group Offloading (#10503)

  Show description for 9a147b8
  

  a-r-r-o-w

  and

  stevhliu

  authored
  9a147b8

  Copy full SHA for 9a147b8

Commits on Feb 13, 2025

• Refactor CogVideoX transformer forward (#10789)

  Show description for ab42820
  a-r-r-o-w

  authored
  ab42820

  Copy full SHA for ab42820

• Update FlowMatch docstrings to mention correct output classes (#10788)

  Show description for 8d081de
  a-r-r-o-w

  authored
  8d081de

  Copy full SHA for 8d081de

• Disable PEFT input autocast when using fp8 layerwise casting (#10685)

  Show description for a0c2299
  a-r-r-o-w

  authored
  a0c2299

  Copy full SHA for a0c2299

• make tensors contiguous before passing to safetensors (#10761)

  Show description for 97abdd2
  faaany

  authored
  97abdd2

  Copy full SHA for 97abdd2

Commits on Feb 12, 2025

• `MultiControlNetUnionModel` on SDXL (#10747)

  Show description for 5105b5a
  

  guiyrt

  and

  hlky

  authored
  5105b5a

  Copy full SHA for 5105b5a

• Fix `use_lu_lambdas` and `use_karras_sigmas` with `beta_schedule=squaredcos_cap_v2` in `DPMSolverMultistepScheduler` (#10740)

  hlky

  authored
  ca6330d

  Copy full SHA for ca6330d

• [Single File] Add Single File support for Lumina Image 2.0 Transformer (#10781)

  Show description for 28f48f4
  DN6

  authored
  28f48f4

  Copy full SHA for 28f48f4

• Faster set_adapters (#10777)

  Show description for 067eab1
  

  Luvata

  and

  sayakpaul

  authored
  067eab1

  Copy full SHA for 067eab1

• Refactor OmniGen (#10771)

  Show description for 57ac673
  

  4 peopleauthored
  57ac673

  Copy full SHA for 57ac673

Commits on Feb 11, 2025

• Add support for lumina2 (#10642)

  Show description for 81440fd
  

  5 peopleauthored
  81440fd

  Copy full SHA for 81440fd

• Add OmniGen (#10148)

  Show description for 798e171
  

  4 peopleauthored
  798e171

  Copy full SHA for 798e171

• speedup hunyuan encoder causal mask generation (#10764)

  Show description for 8ae8008
  dabeschte

  authored
  8ae8008

  Copy full SHA for 8ae8008

Commits on Feb 10, 2025

• Add `Self` type hint to `ModelMixin`'s `from_pretrained` (#10742)

  hlky

  authored
  7fb481f

  Copy full SHA for 7fb481f

• [LoRA] fix peft state dict parsing (#10532)

  Show description for 9f5ad1d
  sayakpaul

  authored
  9f5ad1d

  Copy full SHA for 9f5ad1d

Commits on Feb 7, 2025

• EDMEulerScheduler accept sigmas, add final_sigmas_type (#10734)

  hlky

  authored
  464374f

  Copy full SHA for 464374f

Commits on Feb 6, 2025

• Quantized Flux with IP-Adapter (#10728)

  hlky

  authored
  d43ce14

  Copy full SHA for d43ce14

Commits on Feb 5, 2025

• add provider_options in from_pretrained (#10719)

  Show description for 23bc56a
  

  xieofxie

  and

  hualxie

  authored
  23bc56a

  Copy full SHA for 23bc56a

Commits on Feb 4, 2025

• [Fix] Type Hint in from_pretrained() to Ensure Correct Type Inference (#10714)

  Show description for 5b1dcd1
  

  SahilCarterr

  and

  hlky

  authored
  5b1dcd1

  Copy full SHA for 5b1dcd1

• [bitsandbytes] Simplify bnb int8 dequant (#10401)

  Show description for 5e8e6cb
  sayakpaul

  authored
  5e8e6cb

  Copy full SHA for 5e8e6cb

Commits on Feb 1, 2025

• feat(training-utils): support device and dtype params in compute_density_for_timestep_sampling (#10699)

  Show description for 9f28f1a
  

  badayvedat

  and

  sayakpaul

  authored
  9f28f1a

  Copy full SHA for 9f28f1a

Commits on Jan 31, 2025

• Fix enable memory efficient attention on ROCm (#10564)

  Show description for 1ae9b05
  tenpercent

  authored
  1ae9b05

  Copy full SHA for 1ae9b05

Commits on Jan 29, 2025

• [FIX] check_inputs function in Auraflow Pipeline (#10678)

  Show description for aad69ac
  SahilCarterr

  authored
  aad69ac

  Copy full SHA for aad69ac

• fix(hunyuan-video): typo in height and width input check (#10684)

  badayvedat

  authored
  ea76880

  Copy full SHA for ea76880

• support StableDiffusionAdapterPipeline.from_single_file (#10552)

  Show description for 33f9361
  

  Teriks

  and

  hlky

  authored
  33f9361

  Copy full SHA for 33f9361

Commits on Jan 28, 2025

• [Tests] conditionally check `fp8_e4m3_bf16_max_memory < fp8_e4m3_fp32_max_memory` (#10669)

  Show description for 7b100ce
  sayakpaul

  authored
  7b100ce

  Copy full SHA for 7b100ce

• Refactor gradient checkpointing (#10611)

  Show description for c4d4ac2
  a-r-r-o-w

  authored
  c4d4ac2

  Copy full SHA for c4d4ac2

• [fix] refer use_framewise_encoding on AutoencoderKLHunyuanVideo._encode (#10600)

  Show description for f295e2e
  

  hanchchch

  and

  a-r-r-o-w

  authored
  f295e2e

  Copy full SHA for f295e2e

Commits on Jan 27, 2025

• [core] Pyramid Attention Broadcast (#9562)

  Show description for 658e24e
  

  4 peopleauthored
  658e24e

  Copy full SHA for 658e24e