Currently it's using Llama 2, 16 Billion parameter model. It's already extremely smart, we just have to fine tune it with nostr related data.

https://ai.meta.com/llama/

Reply to this note

Please Login to reply.

Discussion

No replies yet.