What types of models does Arize support?
Arize supports a wide range of model types including classification, regression, ranking, and recommendation models built with popular frameworks such as TensorFlow, PyTorch, scikit-learn, and more.
How does Arize detect data drift?
Arize uses statistical techniques to compare distributions of incoming data against baseline training data, identifying shifts in feature values or relationships that may degrade model performance.
Can Arize be used for on-premises deployments?
Currently, Arize is offered as a cloud-based SaaS platform and does not support on-premises deployment, focusing on ease of integration and scalability in cloud environments.
Is there a free tier available?
Yes, Arize offers a free Starter tier that includes basic monitoring features and supports up to 1 million predictions per month, ideal for individual developers and small teams.