matt-c1/llama-3-quant-comparison

Source code to reproduct the results

Closed this issue · 1 comments

Will you make available the source code to reproduce the results?

The important bits are here: https://github.com/matt-c1/llama-3-quant-comparison?tab=readme-ov-file#inference-code
Make sure to open the folded sections! The code is "hidden" there.

But I don't want to publish the whole script because it's a hacky mess where I kept experimenting with various options.