Pull Requests huggingface/diffusers

Fix for loading of Kohya's Musubi tuner's Flux.2 dev lora

#13189 opened 2026-02-26 18:24 by christopher5106

Fix mutable default arguments, resource leak, and redundant assignment

#13185 opened 2026-02-25 22:19 by jashshah999

[attention backends] fix ring CP for flash and flash 3

#13182 opened 2026-02-24 11:39 by sayakpaul

tests: add regression test for logger NameError in torchao_quantizer

#13175 opened 2026-02-23 10:03 by dhruvildarji

[tests] add attention backend tests.

#13174 opened 2026-02-23 06:01 by sayakpaul

Fix CogVideoX dynamic CFG to use loop index instead of timestep value

#13172 opened 2026-02-23 04:17 by veeceey

[attention backends] remove non-hub attention backends.

#13171 opened 2026-02-22 06:58 by sayakpaul

[tests] refactor UNet model tests to align with the new pattern tests size/L

#13153 opened 2026-02-16 10:16 by sayakpaul

Allows DDPM Scheduler to have Zero Variance on Generation Runs

#13141 opened 2026-02-13 18:20 by rustadadam

Add Claude Diffusers skill and Claude Code plugin scaffold (#12935)

#13138 opened 2026-02-13 16:19 by aazeeem11

fix: freeze shared transformer modules in ZImageControlNet.from_transformer

#13136 opened 2026-02-13 11:56 by jscaldwell55

Fix missing .unsqueeze(1) in attention mask for HunyuanVideo VAE

#13133 opened 2026-02-13 03:34 by Minisal

Fix CogVideoX scheduler prev_timestep for non-leading spacing

#13125 opened 2026-02-11 18:38 by Mr-Neutr0n

Fix: handle remaining meta tensors in from_single_file before dispatch

#13117 opened 2026-02-10 14:26 by kitsune-hash

[tests] tests for `modules_to_not_convert`

#13116 opened 2026-02-10 06:37 by sayakpaul

Fix AudioLDM2Pipeline AttributeError with GPT2Model

#13111 opened 2026-02-09 18:41 by Mr-Neutr0n

Add bounds check for AuraFlow positional embedding indices

#13110 opened 2026-02-09 18:35 by Mr-Neutr0n

Fix SanaCombinedTimestepGuidanceEmbeddings to handle missing guidance

#13109 opened 2026-02-09 18:33 by Mr-Neutr0n

Fix DDIMScheduler.step() prev_timestep for non-leading spacing

#13108 opened 2026-02-09 18:31 by Mr-Neutr0n

fix(pipeline): Preserve dtype in from_pipe() instead of defaulting to float32

#13103 opened 2026-02-08 12:56 by Mr-Neutr0n

fix(controlnet): Use deep copy in ZImageControlNet.from_transformer

#13102 opened 2026-02-08 12:14 by Mr-Neutr0n

Add JiT model and pipeline

#13098 opened 2026-02-08 08:19 by AlanPonnachan

feat(qwenimage): add modular area composition denoise route

#13097 opened 2026-02-08 07:25 by yaoqih

Add ACE-Step pipeline for text-to-music generation

#13095 opened 2026-02-07 11:24 by ChuxiJ

fix: correct Z-Image guidance scale threshold from > 1 to > 0

#13094 opened 2026-02-07 08:08 by Mr-Neutr0n

fix: add PyTorch version check for torch.accelerator in gather_size_by_comm

#13093 opened 2026-02-07 08:06 by Mr-Neutr0n

Add NAG guider and update QwenImage modular pipeline

#13089 opened 2026-02-06 13:40 by yaoqih

Allow `VideoProcessor` to Accept Single Image Inputs

#13084 opened 2026-02-05 02:50 by dg845

[Context Parallel] fix: support PyTorch < 2.6 in distributed gather

#13075 opened 2026-02-03 20:20 by akshatvishu

Add FLUX.2 Klein Inpaint Pipeline close-to-merge tests size/L utils pipelines

#13050 opened 2026-01-28 19:22 by adi776borate