Inference

What is Inference?

Inference is the phase where a trained model is put to work, processing live data to make predictions or generate content.

Where did the term "Inference" come from?

Contrasts with the 'training' phase.

How is "Inference" used today?

The operational cost of AI is often measured in inference costs.

Related Terms

• training
• latency
• Large Language Model (LLM)