downloading and running llama2-7B chat on linux debian
Here are the steps I took to download and run my llama2 7B chat model:
-
Go to https://ai.meta.com/resources/models-and-libraries/llama-downloads/ to request a license to download. You should soon get an email with the download link.
-
Go to your desired folder and git clone the repo:
git clone https://github.com/facebookresearch/llama -
Go to the cloned repo
cd llama -
Runt the download.sh file from terminal
bash ./download.sh -
It should prompt you to paste link. Now paste the link in your email here that begins with https://download.llamameta.net.....
-
It should then prompt you to choose the model. I entered 7b-chat
Enter the list of models to download without spaces (7B,13B,70B,7B-chat,13B-chat,70B-chat), or press Enter for all: 7B-chat -
The process should now begin. It took me almost 10 minutes to download the model.