Skip to content

A straightforward method for training your LLM, from downloading data to generating text.

Commits on main

dec0553
feat(scripts): run every script straight from the repo
FareedKhan-dev committed 1 month ago
2a1efb8
fix(models): forward_embedding with more than one block
FareedKhan-dev committed 1 month ago
83f0b07
feat(inference): shared sampling helpers with min-p
FareedKhan-dev committed 1 month ago
5b2fe83
feat(device): pick CUDA, Apple MPS or CPU automatically
FareedKhan-dev committed 1 month ago
d69b5c2
d58d5f1
6cd3f8d
b995104
docs: rewrite README as a complete end-to-end tutorial
FareedKhan-dev committed 5 months ago
d1b71e0
040f478
0cc6125
fix(ddp): make the reward model train on multiple GPUs
FareedKhan-dev committed 5 months ago
9476be5
806659c
2f995f8
e851928
Add link to Pretraining at bottom.
Mr. Ohu committed 5 months ago
2b59fac
Correct equation format in objectives.md
Mr. Ohu committed 5 months ago
cebcc2d
docs: add foundations tutorial
Abellegese committed 5 months ago
6efc062