this post was submitted on 21 Nov 2023
1 points (100.0% liked)

LocalLLaMA

1 readers
1 users here now

Community to discuss about Llama, the family of large language models created by Meta AI.

founded 10 months ago
MODERATORS
 

Looking for speed and accuracy. Any suggestions on cloud hosts?

you are viewing a single comment's thread
view the rest of the comments
[–] CircumventThisReddit@alien.top 1 points 10 months ago (1 children)

Write your own parser and integrate function calling with any LLm your heart desires.

[–] _nembery@alien.top 1 points 10 months ago

It’s not even that hard. Just use a regex on the return text for simple classification tasks. Any llama2 can do this reasonably well. The hard part is when you want complex JSON data structures