in the blog post you said "[...] mixture of full fine-tuning and LoRA was used to provide better generalization.".could you expand?
· Sign up or log in to comment
