install.sh

#!/bin/bash
set -e

# Check if miniconda3 is installed
if [ ! -d "miniconda3" ]; then
    echo "miniconda3 directory not found, please run the init.sh script first"
    exit 1
fi

# Source Conda shell script
source miniconda3/etc/profile.d/conda.sh

# Create Conda environment
conda create -y -n textgen python=3.10.9 || { echo "Error creating Conda environment"; exit 1; }

# Activate Conda environment
conda activate textgen

# Remove existing apt-key if present
sudo apt-key del 7fa2af80 || true

# Download and install NVIDIA CUDA Toolkit
wget -q --show-progress https://developer.download.nvidia.com/compute/cuda/repos/wsl-ubuntu/x86_64/cuda-wsl-ubuntu.pin
sudo mv cuda-wsl-ubuntu.pin /etc/apt/preferences.d/cuda-repository-pin-600
wget -q --show-progress https://developer.download.nvidia.com/compute/cuda/11.7.0/local_installers/cuda-repo-wsl-ubuntu-11-7-local_11.7.0-1_amd64.deb
sudo dpkg -i cuda-repo-wsl-ubuntu-11-7-local_11.7.0-1_amd64.deb
sudo cp /var/cuda-repo-wsl-ubuntu-11-7-local/cuda-*-keyring.gpg /usr/share/keyrings/
sudo apt-get update
sudo apt-get -y install cuda

# Install required Conda packages
conda install -y -n base conda-libmamba-solver
conda config --set solver libmamba
conda install -y cudatoolkit
conda install -y pytorch torchvision torchaudio pytorch-cuda=11.7 -c pytorch -c nvidia
conda install -y -c conda-forge cudatoolkit=11.7
conda install -y -c conda-forge ninja
conda install -y -c conda-forge accelerate
conda install -y -c conda-forge sentencepiece
conda install -y -c conda-forge gradio
conda install -y -c conda-forge cudatoolkit-dev
conda install -y markdown chardet cchardet ninja

# Fix symlinks for libcuda
sudo rm -f /usr/lib/wsl/lib/libcuda.so /usr/lib/wsl/lib/libcuda.so.1
sudo ln -s /usr/lib/wsl/lib/libcuda.so.1.1 /usr/lib/wsl/lib/libcuda.so.1
sudo ln -s /usr/lib/wsl/lib/libcuda.so.1 /usr/lib/wsl/lib/libcuda.so

# Clone repositories
git clone https://github.com/oobabooga/text-generation-webui
cd text-generation-webui
git reset --hard 8134c4b334ecc3a6a30e774a7265cbafc42ac6cc
pip install -r requirements.txt
mkdir -p repositories
cd repositories
git clone https://github.com/qwopqwop200/GPTQ-for-LLaMa
cd GPTQ-for-LLaMa
git reset --hard 468c47c01b4fe370616747b6d69a2d3f48bab5e4
pip install -r requirements.txt

# Install CUDA version of GPTQ-for-LLaMa
MAX_JOBS=4
python setup_cuda.py install

cd ~/text-generation-webui/
python download-model.py --text-only decapoda-research/llama-7b-hf
python download-model.py decapoda-research/llama-7b-hf-int4
mv models/llama-7b-hf-int4/llama-7b-4bit.pt models/

# Replace LLaMATokenizer with LlamaTokenizer
perl -i -pe 's/LLaMATokenizer/LlamaTokenizer/g' models/llama-7b-hf/tokenizer_config.json

echo "======================================================================"
echo "DONE! Run the following commands and your new friend will be awaiting for you:"
echo "cd text-generation-webui/"
echo "conda activate textgen"
echo "python ./server.py --model llama-7b-hf --cai-chat --gptq-bits 4 --listen"