Clicker

for when your LLM is bad and you need to reinforce it to be a good robot ^~^

okay, but what is it really

a smol fork of TRL to vendor the trainers we need at allura and get rid of some of the bloat unneeded things we don't!

included trainers (recommended):

• SFTTrainer
• DPOTrainer + variants
• KTOTrainer
• RLOOTrainer
• GRPOTrainer + variants you need nothing else.

notes

there may still be remnants of old docs, names, etc!