Model weights
The full set of numbers that make up a trained model. Download the weights and you have the model.
A model is a huge set of dials, each holding a number. Training turns the dials a little at a time, over millions of examples, until the model’s guesses are good. After training the dials are locked: using the model does not change them, and showing it your document does not teach it anything permanent.
The fuller definition
Model weights are learned numerical values that influence how a neural network transforms inputs. Saved weight files hold much of the learned state needed to run the model.
Go deeper
Idealized value storage only, in decimal GB. Metadata, caches, mixed precision, and runtime overhead add to actual requirements. This is not a GPU purchasing calculator.
A concrete example
Downloading model weights lets compatible software load their values for inference, subject to hardware and licensing requirements.
The distinction that matters
Weights are one part of a usable model. You also need the matching configuration, tokenizer where applicable, and a compatible runtime.
Related terms
Parameter
One of the billions of adjustable numbers inside a model. Training sets them; using the model does not change them.
Hardware & hostingOpen weights
A model whose trained numbers are published, so anyone can download and run it.
Hardware & hostingLocal model
A model that runs on your own computer, so your documents never leave it.
Where it is taught
What actually happens when you ask AI a question?Source and further reading
Google · Machine Learning GlossaryThe definition and example are written for this library. The link gives background, official documentation or legal authority as labelled. Legal scope depends on the jurisdiction and current text; product features depend on release and settings.