fp16 is overkill though. 8-bit is the sweet spot before perf degradation starts ... | Hacker News

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

		int_19h 11 months ago \| parent \| context \| favorite \| on: Qwen3: Think deeper, act faster fp16 is overkill though. 8-bit is the sweet spot before perf degradation starts getting noticeable.

rahimnathwani 11 months ago [–]

I haven't yet seen any evals comparing the original Qwen3-30B-A22B with https://ollama.com/library/qwen3:30b-a3b-q8_0

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