Its smart. Definitely keeping it around as my slow/low context local llm.
Seems like a great candiate to merge with other 70Bs as well. There aren't a lot of really great 70B training continuations, like CodeLlama or sequelbox's continuations.
Seems like a great candiate to merge with other 70Bs as well. There aren't a lot of really great 70B training continuations, like CodeLlama or sequelbox's continuations.
No comments yet.