I want to recover the performance of the pruned model using full fine-tuning. But i think that only the LoRA code is realeased. Is there any plan to release the full fine-tuning code? Or can you explain me the protocol about that?
I want to recover the performance of the pruned model using full fine-tuning.
But i think that only the LoRA code is realeased.
Is there any plan to release the full fine-tuning code?
Or can you explain me the protocol about that?