huggingface / trl

Train transformer language models with reinforcement learning.
http://hf.co/docs/trl
Apache License 2.0
10.06k stars 1.27k forks source link

How to train from scratch? Can you provide the code #2356

Open sankexin opened 2 days ago

sankexin commented 2 days ago

System Info

train from scratch

Information

Tasks

Reproduction

train from scratch

Expected behavior

train from scratch

Checklist

qgallouedec commented 2 days ago

Does "from scratch" means the opposite of "finetuning" for you? Please precise your question

kalocide commented 21 hours ago

why would you pre-train with RL?