- CUDA 11.8.0
- Ubuntu 20.04
conda create -n robomonkey python=3.10 -y
conda activate robomonkeygit clone https://github.com/robomonkey-vla/monkey-verifier
cd monkey-verifier/llava_setup
git clone https://github.com/robomonkey-vla/LLaVA
cd LLaVA
git apply < ../fix_llava_padding.patch
pip install --upgrade pip # enable PEP 660 support
pip install -e .
pip install ninjapip install torch==2.0.1+cu118 torchvision==0.15.2+cu118 torchaudio==2.0.2 --index-url https://download.pytorch.org/whl/cu118
pip install deepspeed==0.9.3
pip install peft==0.4.0
pip install transformers==4.31.0
pip install bitsandbytes==0.41.0
pip install datasets
pip install wandb
pip install "numpy<2"cd ~/monkey-verifier/model_dir
git clone https://huggingface.co/zhiqings/LLaVA-RLHF-7b-v1.5-224cd ~/monkey-verifier/data_dir
git clone https://huggingface.co/datasets/robomonkey-vla/bridge_data_v2
unzip bridge_v2_images.zip
cd ~/monkey-verifier/data_dir
git clone https://huggingface.co/datasets/robomonkey-vla/action_preference_bridgeRun the training script:
cd ~/monkey-verifier/RLHF
bash scripts/7b-v1.5-224/full_bridge_finetune.shBuilt on LLaVA-RLHF by Sun et al.