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

LocalLLaMA

3 readers
1 users here now

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

founded 1 year 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 11 months ago (1 children)

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

[–] _nembery@alien.top 1 points 11 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