login sign up

80% faster, 50% less memory, 0% loss of accuracy Llama finetuning

github.com · first added by @hn_wayback
saved by
1 person
discussions · 1
Show HN: 80% faster, 50% less memory, 0% loss of accuracy Llama finetuning 385 pts · 119 comments · Dec 2023 385 pts · 119 comments · Dec 2023
0 people saved or upvoted this