Why was finetune deleted?

[mboros1]

I was investigating the possibility of creating LoRA's with llama.cpp, and saw that it used to exist with the finetune tool but was deleted earlier this year, be6d7c0

Does anyone have insights on why it was removed?

[hieuhthh]

After quantize, my generated result is worse. I want to improve this quantized model, like through calibartion, lora, finetune, etc ... Any suggestion or solutions please?