instructions.txt

mkdir -p ~/miniconda3
wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh -O ~/miniconda3/miniconda.sh
bash ~/miniconda3/miniconda.sh -b -u -p ~/miniconda3
rm -rf ~/miniconda3/miniconda.sh


git clone https://github.com/Vision-CAIR/MiniGPT-4.git
cd MiniGPT-4
conda env create -f environment.yml
conda activate minigpt4

Install git lfs
git lfs install
git clone https://huggingface.co/Vision-CAIR/vicuna

minigpt4/configs/models/minigpt4_vicuna0.yaml Change path of llama_model in line 18 to the above downloaded folder directory

pip install gdown
gdown --id 1RY9jV0dyqLX-o38LrumkKRh6Jtaop58R -O prerained_minigpt4_7b.pth

eval_configs/minigpt4_eval.yaml Change path to above file in line 8

You are all set to run the demo file using this command

python demo.py --cfg-path eval_configs/minigpt4_eval.yaml  --gpu-id 0

Verify if its working till now.

Copy paster cloud_api_minigpt.py file from our repo
and then MiniGPT-4/minigpt4/common/config.py Change Line 25 and Line 27 with this

user_config = self._build_opt_list("+")
 
config = OmegaConf.load('eval_configs/minigpt4_eval.yaml')

Run the server with this, save the file as test_Api.py, run uvicorn test_api:app --port 8080 --reload

Now your minigpt is hosted locally

create a ngrok acc and host it globally
Its done! Woooohoooo!