huggingface/diffusers

Pull Requests Commits

fix: simplify with latest kernels private apis

drbh committed 2 days ago

Verified cc4003b2

fix: pin v2 and wrap ops like fa3

drbh committed 2 days ago

Verified 6d80e29b

fix: prefer flash attn2 public api

drbh committed 2 days ago

Verified bd1a2950

[TPU] TorchTPU backend integration - eager / torch.compile / tp (#14039)

JingyaHuang committed 2 days ago

Verified 1d5d056e

[LoRA] add LoKr adapter support (Z-Image, Flux2/Klein) (#14163)

christopher5106 committed 2 days ago

Verified 78bf2c62

[ci] consolidate labeling on PRs. (#14911)

sayakpaul committed 2 days ago

Verified d7d0b9c2

fix(image_processor): maintain 3-tuple return contract in InpaintProcessor.preprocess when mask is None (#14481)

patrickswedish committed 3 days ago

Verified d961a388

Fix active adapter detection for text encoders (#14961)

Subodh-17 committed 3 days ago

Verified e0415c41

[docs] Community methods (#14973)

stevhliu committed 4 days ago

Verified 122b1e11

[docs] Jobs guide (#14926)

stevhliu committed 5 days ago

Verified c6df88a5

Bump doc-builder workflow pin (#14957)

paulinebm committed 5 days ago

Verified 71774d9b

Kandinsky6 (#14949)

sayakpaul committed 5 days ago

Verified bc5e3bd9

fix https://github.com/huggingface/diffusers/issues/14461 (#14956)

yiyixuxu committed 5 days ago

Verified 36438e2e

[tests] remove `test_pipelines_common.py` (#14669)

sayakpaul committed 5 days ago

Verified 77177984

[chore] remove force_upcast from autoencoder flux2 (#14946)

sayakpaul committed 5 days ago

Verified 899c9f3f

Configure fixed sampling sigmas on QwenImage21Pipeline (#14950)

naykun committed 6 days ago

Verified da1d3829

[tests] refactor the pipeline-level attention processor tests (#14701)

sayakpaul committed 6 days ago

Verified c2798cc7

[modular] Echo team joy future academy jd echo by @Gelercatty (#14941)

sayakpaul committed 6 days ago

Verified cff9dafb

Fix: Deprecating Siglip2ImageProcessorFast (#14781)

ariG23498 committed 6 days ago

Verified 9aababd7

agent docs: support only what released checkpoints use (#14750)

yiyixuxu committed 9 days ago

Verified 8b33bfc0

[Bugfix] Fix Cosmos3 FP8 tensor parallelism (#14921)

yzhautouskay committed 9 days ago

Verified 357f8738

Tag Krea 2 text encoder outputs as denoiser_input_fields (#14925)

tbsbngrtz committed 9 days ago

Verified 41566305

Add single file support for Krea 2 (#14914)

DN6 committed 10 days ago

Verified 578c9b2c

chore: add one additional copied from in qwenimage 2.1 vae (#14810)

sayakpaul committed 10 days ago

Verified 7997e4ea

Consolidate torch device backend dispatch (#14792)

DN6 committed 10 days ago

Verified acbabca3

Add single file support for Minimax H3 (#14839)

DN6 committed 10 days ago

Verified 4de185d6

[Modular] Avoid downloading weights when loading from an existing local path (#14797)

DN6 committed 10 days ago

Verified a91ff967

[core] Shard tensor-parallel checkpoints on load and save (#14544)

JingyaHuang committed 10 days ago

Verified 863092ff

[fix] Add return types (#14874)

stevhliu committed 11 days ago

Verified c60830ee

[Cosmos3] Fix Transfer SeaCache artifacts with control CFG (#14897)

yzhautouskay committed 11 days ago

Verified 78405ec9

Older