princeton-nlp / SWE-agent

SWE-agent takes a GitHub issue and tries to automatically fix it, using GPT-4, or your LM of choice. It solves 12.47% of bugs in the SWE-bench evaluation set and takes just 1 minute to run.
https://princeton-nlp.github.io/SWE-agent/
MIT License
11.86k stars 1.18k forks source link

Add claude 3.5 sonnet to models #601

Closed carlosejimenez closed 1 week ago

carlosejimenez commented 1 week ago

Add claude 3.5 sonnet to models and add a default_xml config to configs to better support claude models

Reference Issues/PRs

Fixes #600

What does this implement/fix? Explain your changes.

Adds model name and cost details to sweagent/agent/models.py Also adds the default_xml.yaml in configs, which is the preferred version to use with claude.

codecov[bot] commented 1 week ago

Codecov Report

All modified and coverable lines are covered by tests :white_check_mark:

Project coverage is 76.40%. Comparing base (b6257de) to head (03b9339).

Additional details and impacted files ```diff @@ Coverage Diff @@ ## main #601 +/- ## ======================================= Coverage 76.40% 76.40% ======================================= Files 23 23 Lines 3387 3387 ======================================= Hits 2588 2588 Misses 799 799 ```

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.

klieret commented 1 week ago

Awesome 😃

EwoutH commented 1 week ago

Amazing how fast this was added!

Could a new point release be tagged, so we can install and use it easily?

klieret commented 1 week ago

@EwoutH Until we have a proper python API and meaningfully release to pypi, releases are purely informational (since the installation instructions are just off of main). So it's probably gonna part of v0.7.0 (since I just released v0.6.1 before)