diffusers
https://github.com/huggingface/diffusers
Python
🤗 Diffusers: State-of-the-art diffusion models for image, video, and audio generation in PyTorch.
Triage Issues!
When you volunteer to triage issues, you'll receive an email each day with a link to an open issue that needs help in this project. You'll also receive instructions on how to triage issues.
Triage Docs!
Receive a documented method or class from your favorite GitHub repos in your inbox every day. If you're really pro, receive undocumented methods or classes and supercharge your commit history.
Python not yet supported3 Subscribers
Add a CodeTriage badge to diffusers
Help out
- Issues
- QwenImage21 FlexAttention mask fails to compile on MPS because of boolean ~
- Converted sigma schedules with final_sigmas_type="sigma_min" repeat the last sigma, and the third-order and heun updates of DPMSolverMultistepScheduler and UniPCMultistepScheduler return NaN on that zero-width final step
- `Cosmos2VideoToWorldPipeline` conditions on the entire input video — no `num_conditional_frames`, so a full clip is reconstructed instead of continued
- fix(dreambooth): align Flux2 Klein text_encoder_out_layers default with pipeline
- Legacy `final_offload_hook` checks are dead in 14 pipelines; `KandinskyPriorPipeline` skips `maybe_free_model_hooks()` when `negative_prompt` is set
- `apply_group_offloading(use_stream=True)` raises on Ascend NPU: no stream branch for `torch.npu`
- `Pipeline.to()` offload guards only cover `cuda`/`xpu`: Ascend NPU misses the sequential-offload conflict error and the model-offload warning
- Five Z-Image pipelines miss the precomputed-t_norms DtoH-sync fix from ZImagePipeline
- Cosmos denoising loops sync with the host every step
- WanAnimatePipeline.get_i2v_mask() defaults device to "cuda", which raises on non-CUDA accelerators (NPU/XPU/MPS)
- Docs
- Python not yet supported