Eval infra + env fixes + new notebook params

Azure-Samples / raft-distillation-recipe

A recipe that will walk you through using either Meta Llama 3.1 405B or GPT-4o deployed on Azure AI to generate a synthetic dataset using UC Berkeley's Gorilla project RAFT method.

MIT License

36 stars 11 forks source link

Eval infra + env fixes + new notebook params #34

Closed cedricvidal closed 2 months ago

cedricvidal commented 2 months ago

finetuning_threshold is now a notebook parameter
azd env should override state env
Provisioning scoring model
Introduced parameter raft_questions
Rely on previously exported .env.state file