HackerTrans
TopNewTrendsCommentsPastAskShowJobs

minosu

no profile record

Submissions

Simple GRPO – RL for 8B models on $10/h GPUs

github.com
1 points·by minosu·last year·0 comments

Show HN: AIWP – Context-aware DALL·E 3 wallpapers for Android

github.com
8 points·by minosu·3 years ago·0 comments

comments

minosu
·2 years ago·discuss
Mistral 7b inferences about 18% faster for me as a 4bit quantized version on an A100. Thats definitely relevant when running anything but chatbots.