requirements.txt

torch>=2.2.2
torchvision>=0.17.2
torchaudio>=2.2.2
transformers==4.51.0
diffusers==0.33.0
decord>=0.6.0
attrdict2
accelerate>=0.32.0
timm>=1.0.15
opencv-python>=4.10.0
pillow>=10.4.0
einops>=0.8.0
xformers>=0.0.28
numpy>=1.26.0
pandas>=2.2.0
datasets>=3.0.0
tokenizers>=0.21.0
sentencepiece>=0.1.99
torchmetrics>=1.4.0
tqdm>=4.66.0
pyyaml>=6.0.0
requests>=2.32.0
packaging>=24.1
ipython>=8.26.0
matplotlib>=3.9.0
deepspeed>=0.14.4  
wandb>=0.16.3      
gradio>=5.34.0    
qwen-vl-utils