# On Windows, it is difficult to prepare flash_attn2 and probably cannot run. # On WSL2: # sudo apt install python3.10-dev # wget https://developer.download.nvidia.com/compute/cuda/repos/wsl-ubuntu/x86_64/cuda-keyring_1.0-1_all.deb # sudo dpkg -i cuda-keyring_1.0-1_all.deb # sudo apt update # sudo apt-get install cuda-toolkit-12-1 # vi ~/.bashrc # if [ -e /usr/local/cuda ]; then # export PATH="/usr/local/cuda/bin:$PATH" # export LD_LIBRARY_PATH="/usr/local/cuda/lib64:$LD_LIBRARY_PATH" # fi fastapi uvicorn transformers==4.43.3 bitsandbytes==0.43.3 accelerate==0.33.0 peft==0.12.0 wheel python-multipart -r requirements-torch.txt -r requirements-fa.txt unsloth[cu121-ampere-torch230] @ git+https://github.com/unslothai/unsloth.git"