This is my repo for experimenting with Meta Llama, a free language model.
Follow this webpage to download the original model weights (~14 GB). Basically, the weights in this repo are only a small subset of the total weights, so you need the original model, which you then add the fine-tuned weights from each subdirectory onto.