cd /nfs/ywang29/video_reward/DanceGRPO

source activate 
# conda create -n wan python==3.10 -y
# conda activate wan

./env_setup.sh fastvideo


pip3 install moviepy
pip3 install protobuf==3.20.0
# mkdir videos
pip3 install huggingface_hub==0.24.0 
pip3 install tf-keras
pip3 install trl
# pip3 install transformers==4.51.0

pip install torchvision==0.21.0
pip install pydantic -U
pip install qwen_vl_utils
pip install librosa
pip install wandb


cd /nfs/ywang29/Reward_finetuning/VideoX-Fun
export LD_LIBRARY_PATH=/usr/local/cuda/lib64

# pip install -U accelerator
pip install transformers==4.57
pip install accelerate -U
pip install flash-attn==2.5.8 --no-build-isolation
# cp /nfs/ywang29/Reward_finetuning/VideoX-Fun/scripts/wan2.1/modeling_qwen2_5_vl.py /opt/conda/lib/python3.10/site-packages/transformers/models/qwen2_5_vl/modeling_qwen2_5_vl.py