requirements.txt

# Core dependencies
numpy
einops
librosa
pydub
torch
torchaudio
torchvision
torchcodec
torchdiffeq
transformers>=4.54.0

# Audio / multimodal models
audiobox_aesthetics

# Git-based dependencies
git+https://github.com/huggingface/datasets.git
git+https://github.com/huggingface/peft.git
git+https://github.com/facebookresearch/sam3.git
git+https://github.com/huggingface/accelerate.git
dacvae @ git+https://github.com/facebookresearch/dacvae.git
imagebind @ git+https://github.com/facebookresearch/ImageBind.git
laion-clap @ git+https://github.com/lematt1991/CLAP.git
git+https://github.com/wookayin/nvidia-ml-py.git
#perception-models @ git+https://github.com/facebookresearch/perception_models@unpin-deps
perception-models@git+https://github.com/hx23840/perception_models@unpin-deps