Relevant_Outcome_726

joined 1 year ago

Are 7b models useful? in c/localllama@poweruser.forum

[–] Relevant_Outcome_726@alien.top 1 points 11 months ago

Can you provide the prompt for function call?

permalink
fedilink
source
context

Practical Tips for Finetuning LLMs Using LoRA (Low-Rank Adaptation) in c/localllama@poweruser.forum

[–] Relevant_Outcome_726@alien.top 1 points 1 year ago

From my experience, Here are some other things related to Lora:
+ FSDP doesn't work for Lora because FSDP requires all parameters to be trainable or frozen.

+ For Qlora, currently we can only use deepspeed zero2 (deepspeed zero 3 is not supported)

permalink
fedilink
source