3 ms·
The TinyLlama project is trying to do that pushing by training a small 1.1 billion-parameter model on 3 trillion tokens: https://github.com/jzhang38/TinyLlama h
by yorwba 3y ago
The TinyLlama project is trying to do that pushing by training a small 1.1 billion-parameter model on 3 trillion tokens: https://github.com/jzhang38/TinyLlama https://github.com/jzhang38/TinyLlama