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

communick@poweruser.forum

Structured Output with Zephyr (alien.top)

submitted 10 months ago by Smsmith714@alien.top to c/localllama@poweruser.forum

4 comments fedilink hide all child comments

I've been experimenting with Zephyr and pretty surprised by the great performance. One problem I have with Zephyr is the difficulty to build structured outputs. For example if I ask it to return only True or False, it will return a lengthy explanation. This makes it tough to use Zephyr as part of a production system, since incorrectly structured outputs have huge implications.

Has anyone found some tricks to make Zephyr produce outputs in a defined format and do so deterministically?

top 4 comments

sorted by: hot top controversial new old

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

There are all sorts of approaches

microsoft guidance
llama.cpp grammar constraints
someone recently made their own approach and posted it here, search for CAPPr

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

The correct answer here is reading logprobs/logits over the true/false tokens after parsing the input.

Using a grammar is overkill and a classification model might not get you the fidelity you want.

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

Fine tune of course

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

I reverted to mistral-instruct