The Verge
6 months agoArtificial intelligence
Amazon will offer human benchmarking teams to test AI models
Amazon has launched Model Evaluation on Bedrock, a tool to help developers test the performance of AI models.
The tool includes both automated evaluation and human evaluation components to ensure accurate and reliable model selection.
Users can bring their own data into the platform for benchmarking purposes, and can choose to work with either an AWS human evaluation team or their own. [ more ]