Parameters

Models and Architectures

The internal numbers a model learns during training. Model sizes are described by parameter count, like 7B or 70B (billions).

Parameters are the adjustable numbers inside a neural network - the learned connection strengths that encode everything the model knows. Training is the process of tuning them.When you read that a model has 70B parameters, that means 70 billion of these numbers. More parameters generally means more capability but also more computing power needed, which is why small models run on phones while the largest need data centers.