I've seen 3 epochs on some of the finetuning R1 blog posts. It's not my field so... | Hacker News

Hacker Newsnew | past | comments | ask | show | jobs | submit

		_joel 7 months ago \| parent \| context \| favorite \| on: SmolGPT: A minimal PyTorch implementation for trai... I've seen 3 epochs on some of the finetuning R1 blog posts. It's not my field so not sure how valid that is.

sva_ 7 months ago [–]

Yeah, fine-tuning is different from pretraining

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact