name: diffusers version: 0.40.0 summary: State-of-the-art diffusion in PyTorch. license: Apache Software License author: The Hugging Face team (past and future) with the help of all our contributors (https://github.com/huggingface/diffusers/graphs/contributors) homepage: https://github.com/huggingface/diffusers requires_python: >=3.10.0 python_min: 3.10.0 install_size: 5.6 MB module_format: py3-none-any repo_url: https://github.com/huggingface/diffusers last_released_at: 2026-08-20T14:30:38Z releases_last_12mo: 7 avg_days_between_releases: 15.7 platforms: any ## dependencies importlib_metadata filelock httpx <1.0.0 huggingface-hub <2.0,>=1.23.0 numpy regex !=2019.12.17 requests safetensors >=0.8.0 Pillow ## extras [bitsandbytes] bitsandbytes >=0.43.3 accelerate >=0.31.0 [dev] urllib3 <=2.0.0 isort >=5.5.4 ruff ==0.9.10 hf-doc-builder >=0.3.0 ftfy GitPython <3.1.19 datasets Jinja2 invisible-watermark >=0.2.0 librosa llvmlite >=0.40.0 numba >=0.57.0 parameterized protobuf <4,>=3.20.3 pytest pytest-timeout pytest-xdist requests-mock ==1.10.0 safetensors >=0.8.0 sentencepiece !=0.1.92,>=0.1.91 scipy tiktoken >=0.7.0 torchsde torchvision transformers >=4.41.2 phonemizer accelerate >=0.31.0 datasets protobuf <4,>=3.20.3 tensorboard Jinja2 peft >=0.17.0 timm hf-doc-builder >=0.3.0 torch >=2.6 accelerate >=0.31.0 [docs] hf-doc-builder >=0.3.0 [flashpack] flashpack [gguf] gguf >=0.10.0 accelerate >=0.31.0 [nvidia-modelopt] nvidia_modelopt[hf] >=0.33.1 [optimum-quanto] optimum_quanto >=0.2.6 accelerate >=0.31.0 [quality] urllib3 <=2.0.0 isort >=5.5.4 ruff ==0.9.10 hf-doc-builder >=0.3.0 [sdnq] sdnq >=0.2.2 [test] ftfy GitPython <3.1.19 datasets Jinja2 invisible-watermark >=0.2.0 librosa llvmlite >=0.40.0 numba >=0.57.0 parameterized protobuf <4,>=3.20.3 pytest pytest-timeout pytest-xdist requests-mock ==1.10.0 safetensors >=0.8.0 sentencepiece !=0.1.92,>=0.1.91 scipy tiktoken >=0.7.0 torchsde torchvision transformers >=4.41.2 phonemizer [torch] torch >=2.6 accelerate >=0.31.0 [torchao] torchao >=0.7.0 accelerate >=0.31.0 [training] accelerate >=0.31.0 datasets protobuf <4,>=3.20.3 tensorboard Jinja2 peft >=0.17.0 timm ## maintainers The Hugging Face team (past and future) with the help of all our contributors (https://github.com/huggingface/diffusers/graphs/contributors)