Unsloth creators fix universal error with gradient accumulation | Hacker News

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

		Unsloth creators fix universal error with gradient accumulation (unsloth.ai)
		4 points by ZQ-Dev8 on Oct 16, 2024 \| hide \| past \| favorite \| 2 comments

gnabgib on Oct 16, 2024 | [–]

Article title: Bugs in LLM Training - Gradient Accumulation Fix

ZQ-Dev8 on Oct 16, 2024 | [–]

seems like something pytorch maintainers would want to know about and fix asap...

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