When I try meta-llama/Llama-2-7b-chat-hf in the hf chat, I get very good answers specially for medical coding questions. But when I deployed Sagemaker version of the 70b chat, i get different quality of response.
What can be the cause of this?
When I try meta-llama/Llama-2-7b-chat-hf in the hf chat, I get very good answers specially for medical coding questions. But when I deployed Sagemaker version of the 70b chat, i get different quality of response.
What can be the cause of this?