| # Torch and related libraries | |
| torch | |
| torchvision | |
| torchaudio | |
| xformers | |
| torchao | |
| # Facenet and media processing | |
| facenet_pytorch==2.6.0 | |
| mediapipe | |
| # Transformers, diffusers, and machine learning | |
| transformers>=4.46.3 | |
| diffusers==0.31.0 | |
| torchmetrics | |
| torchtyping | |
| accelerate==1.1.1 | |
| # Image and video processing | |
| opencv-python | |
| opencv-contrib-python==4.10.0.84 | |
| Pillow>=10.2.0 | |
| imageio==2.36.0 | |
| imageio-ffmpeg==0.5.1 | |
| av==13.1.0 | |
| decord==0.6.0 | |
| moviepy==1.0.3 | |
| scenedetect | |
| # NumPy, SciPy, and related libraries | |
| numpy==1.26.4 | |
| scikit-image==0.24.0 | |
| scikit-learn==1.5.2 | |
| scipy==1.14.1 | |
| # Audio processing | |
| soundfile | |
| ffmpeg-python | |
| # Miscellaneous libraries | |
| einops==0.8.0 | |
| omegaconf==2.3.0 | |
| mlflow==2.18.0 | |
| controlnet-aux==0.0.9 | |
| IPython | |
| gradio | |
| gradio_client==1.4.3 | |
| huggingface_hub==0.26.2 | |
| torchdiffeq==0.2.5 | |
| torchsde==0.2.6 | |
| open-clip-torch==2.29.0 | |
| # CLIP library from GitHub | |
| clip @ https://github.com/openai/CLIP/archive/d50d76daa670286dd6cacf3bcd80b5e4823fc8e1.zip#sha256=b5842c25da441d6c581b53a5c60e0c2127ebafe0f746f8e15561a006c6c3be6a | |
| # ONNX Runtime for GPU | |
| onnxruntime-gpu==1.20.1 | |