Custom 4-bit Finetuning 5-7 times faster inference than QLora
#13
pinned
by
rmihaylov
- opened
FalconLLM
pinned discussion
This is amazing, thank you for sharing! We have pinned it so that more people can play with this great library π.