Significant-Gravitas / Auto-GPT-Benchmarks

A repo built for the purpose of benchmarking the performance of agents, regardless of how they are set up and how they work.
MIT License
275 stars 76 forks source link

Use agent protocol for Benchmarks #271

Closed jakubno closed 1 year ago

jakubno commented 1 year ago

Background

Using Agent protocol to run benchmarks. I added downloading and uploading files, this allows to benchmark agents remotely and overall it simplifies the setup.

Other insights are more than welcomed!

Changes

Add option to run benchmarks in API mode, no need to change your agent, if you use agent protocol.

PR Quality Checklist