sabareesh88/fw14k

This is a Llama 2 architecture model series trained on the FineWeb dataset. This is ~500 Million model uses lamma tokenizer. trained using code from Karpathy lamma2