Evals
Quick Start
There are many ways to use Athina Evals.
Running Evals in UI
- Run evals on datasets
- Configure online evals to run automatically: These will run automatically on your logged inferences, and you can view the results in the dashboard.
- Run evaluations manually from the Trace view
- Run evaluations in the Prompt playground
- Configure Automations to run evals automatically on new datasets in a project
Run Evals Programmatically
- Run evals in CI / CD pipelines
- Run evals as guardrails around inference
- Run evals using the Python SDK
Here’s a quickstart guide for running evals using the Python SDK:
1. Install the athina
package
2. Set your API keys
If you are using the python SDK, then can set the API keys like this:
If you are using the CLI, then run athina init
, and enter the API keys when prompted.
3. Load your dataset like this:
You can also load data using a Python Dictionary
4. Now you can run evals like this.
For more details, see this guide on running evals.