Compare your models with Arize experiment results.
Evaluates models on datasets and gives users recorded experiments, evaluation scores, and comparisons in Arize.
When to use it
Use when asked to run experiments, compare models, benchmark performance, or analyze evaluation results.
Give it a model evaluation task; it records experiment outputs and scores in Arize or analyzes existing experiment results.
Requires ax version 0.39.0 or higher and a configured Arize profile; setup instructions cover missing or outdated installations.
Requires authenticated Arize access through a configured profile, using an API key or browser-based OAuth.
This skill
Arize
Uploads experiment outputs and scores (irreversible)
ax profile
Repairs your Arize profile (irreversible)