This repository has been archived on 2024-10-27. You can view files and clone it, but cannot push or open issues or pull requests.
local-llm-server/other/vllm/build-vllm.sh

39 lines
1001 B
Bash
Raw Permalink Normal View History

2023-09-13 11:22:33 -06:00
#!/bin/bash
# Expected to be run as root in some sort of container
cd /tmp || exit
if [ ! -d /tmp/vllm-gptq ]; then
git clone https://github.com/chu-tianxiang/vllm-gptq.git
cd vllm-gptq || exit
else
cd vllm-gptq || exit
git pull
fi
if [ ! -d /root/miniconda3 ]; then
wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh -O Miniconda3-latest-Linux-x86_64.sh
bash /tmp/Miniconda3-latest-Linux-x86_64.sh -b
rm /tmp/Miniconda3-latest-Linux-x86_64.sh
fi
eval "$(/root/miniconda3/bin/conda shell.bash hook)"
if [ ! -d /root/miniconda3/envs/vllm-gptq ]; then
conda create --name vllm-gptq -c conda-forge python=3.11 -y
conda activate vllm-gptq
pip install ninja
conda install -y -c "nvidia/label/cuda-11.8.0" cuda==11.8
conda install -y cudatoolkit cudnn
else
conda activate vllm-gptq
fi
pip install -r requirements.txt
CUDA_HOME=/root/miniconda3/envs/vllm-gptq python setup.py bdist_wheel
echo -e "\n\n===\nOUTPUT:"
find /tmp/vllm-gptq -name '*.whl'