requirements.txt

torch>=2.3.0
torchvision
git+https://github.com/huggingface/transformers@f742a644ca32e65758c3adb36225aef1731bd2a8
accelerate>=0.30.0
qwen-omni-utils[decord]>=0.0.4  # For multimedia processing
soundfile>=0.12.1  # Audio support