huggingface/diffusers

Pull Requests Commits

Merge branch 'main' into save-automodel

DN6 committed 29 days ago

Verified 95daca1d

[AutoModel] Fix bug with subfolders and local model paths when loading custom code (#13197)

DN6 committed 29 days ago

Verified e7fe4ce9

remove db utils from benchmarking (#13199)

sayakpaul committed 29 days ago

Verified 3d908556

[tests] enable cpu offload test in torchao without compilation. (#12704)

sayakpaul committed 29 days ago

Verified 5b544961

[chore] updates in the pypi publication workflow. (#12805)

sayakpaul committed 29 days ago

Verified fcdd759e

[modular] fallback to default_blocks_name when loading base block classes in ModularPipeline (#13193)

yiyixuxu committed 31 days ago

Verified 39188248

[tests] consistency tests for modular index (#13192)

sayakpaul committed 31 days ago

Verified 9b979324

[Modular] update the auto pipeline blocks doc (#13148)

yiyixuxu committed 31 days ago

Verified 680076fc

Fixing Kohya loras loading: Flux.1-dev loras with TE ("lora_te1_" prefix) (#13188)

christopher5106 committed 32 days ago

Verified 5910a1cc

Fix LTX-2 image-to-video generation failure in two stages generation (#13187)

Songrui625 committed 32 days ago

Verified 40e96454

Fix Flash Attention 3 interface for new FA3 return format (#13173)

veeceey committed 33 days ago

Verified 47455bd1

DN6 committed 33 days ago

f20e0f4e

DN6 committed 33 days ago

ba9d2490

Fix wrong `do_classifier_free_guidance` threshold in ZImagePipeline (#13183)

kirillsst committed 33 days ago

Verified 97c2c6e3

Cosmos Transfer2.5 Auto-Regressive Inference Pipeline (#13114)

miguelmartin75 committed 33 days ago

Verified 212db7b9

[attention backends] use dedicated wrappers from fa3 for cp. (#13165)

sayakpaul committed 33 days ago

Verified 31058485

[docs] Fix torchrun command argument order in docs (#13181)

sayakpaul committed 34 days ago

Verified aac94bef

fix: graceful fallback when attention backends fail to import (#13060)

sym-bot committed 35 days ago

Verified 1f6ac1c3

migrate to `transformers` v5 (#12976)

sayakpaul committed 35 days ago

Verified 5e94d62e

Fix AutoModel `typing` Import Error (#13178)

dg845 committed 35 days ago

Verified 7ab20117

Allow Automodel to use `from_config` with custom code. (#13123)

DN6 committed 35 days ago

Verified 4890e9bf

Fix T5GemmaEncoder loading for transformers 5.x composite T5GemmaConfig (#13143)

DavidBert committed 36 days ago

Verified f1e59141

Support Flux Klein peft (fal) lora format (#13169)

asomoza committed 38 days ago

Verified a80b1921

[gguf][torch.compile time] Convert to plain tensor earlier in dequantize_gguf_tensor (#13166)

anijain2305 committed 39 days ago

Verified 01de02e8

[CI] Fix new LoRAHotswap tests (#13163)

DN6 committed 39 days ago

Verified db2d7e7b

remove deps related to test from ci (#13164)

sayakpaul committed 39 days ago

Verified f8d3db9c

[core] Enable CP for kernels-based attention backends (#12812)

sayakpaul committed 40 days ago

Verified 99daaa80

Fix `ftfy` import for PRX Pipeline (#13154)

dg845 committed 40 days ago

Verified fe78a7b7

[CI] Revert `setuptools` CI Fix as the Failing Pipelines are Deprecated (#13149)

dg845 committed 40 days ago

Verified 53e1d0e4

Flux2: Tensor tuples can cause issues for checkpointing (#12777)

dxqb committed 40 days ago

Verified a577ec36

Older