search

FormulaMonks / kurt

A wrapper for AI SDKs, for building LLM-agnostic structured AI applications in Typescript
MIT License
2 stars 0 forks source link

feat: pass KurtSamplingOptions to the underlying LLM provider #32

Closed jemc closed 4 months ago

jemc commented 4 months ago

This is a followup to PR #24, which added KurtSamplingOptions to the Kurt class.

Now, with this PR, the LLM-specific adapters (KurtOpenAI and KurtVertexAI) will now pass these options along to the underlying LLM provider, in the appropriate adapted way.

This commit also adds tests which demonstrate the passing of these options.

theoremonemehdi commented 4 months ago

Tests are failing

jemc commented 4 months ago

Fixed the failing test. It was just a whitespace issue. Fixed and squashed.

github-actions[bot] commented 4 months ago

:tada: This PR is included in version @formula-monks/kurt-open-ai-v1.3.0 :tada:

The release is available on:

Your semantic-release bot :package::rocket:

github-actions[bot] commented 4 months ago

:tada: This PR is included in version @formula-monks/kurt-vertex-ai-v1.2.0 :tada:

The release is available on:

Your semantic-release bot :package::rocket: