Training

Training is the process where an AI model learns from data and adjusts its parameters step by step until it masters the task as well as possible. A distinction is made between pre-training (large, general model from massive data) and fine-tuning (specialization for a task). Training is compute-intensive and often requires many GPUs and weeks to months.

Related Terms

• Machine Learning (ML)
• Fine-Tuning
• Inference
• GPU (Graphics Processing Unit)
• Parameters

Sources: a16z AI Glossary · CNET AI Terms Glossary

← Back to AI Glossary