It matches LLaMA 7B, and it's "cheap" to train for a demo. If they actually want...

		brucethemoose2 on July 6, 2023 \| parent \| context \| favorite \| on: InternLM – new open source 7B LLM It matches LLaMA 7B, and it's "cheap" to train for a demo. If they actually wanted to finetune/train for commodity hardware use, 13b-40b would be a better target.