Training

Training is the process of adjusting a model to improve performance on an objective using data, feedback, or interaction.

Plan for this page

Understand one concept in plain language and connect it to nearby ideas.

You will leave with: A training procedure compares outputs with a target or signal and updates parameters to reduce measured error or increase reward. Choices about examples, objectives, sampling, validation, and stopping shape what the model learns and misses.
Time: 3–8 min
Before you begin: No prerequisite. Keep one real example in mind.
Do this now: Read the explanation, then compare the example with the common misunderstanding.

Prepared next stepCompare training and inference

TrainingTraining is the process of adjusting a model to improve performance on an objective using data, feedback, or interaction.

Model
Parameters
Inference
Dataset
Evaluation

Plain explanation

A training procedure compares outputs with a target or signal and updates parameters to reduce measured error or increase reward. Choices about examples, objectives, sampling, validation, and stopping shape what the model learns and misses.

Example

A classifier is shown labeled plant images, makes predictions, and has its parameters updated when those predictions differ from the labels.

Why it matters

Training and using a model are different stages with different costs, data flows, and controls. Knowing the difference helps evaluate claims about learning from your current conversation.

Common misunderstanding

Misconception: every use of a model retrains it immediately on the user’s input. Correction: inference often uses fixed parameters, while any later data retention or training is a separate product-policy question.

Related concepts

Follow relationships, not only the alphabet

ModelA model is a learned or designed representation that maps inputs to outputs for a defined task or family of tasks.ParametersParameters are internal numerical values that a model’s training procedure adjusts to shape its behavior.InferenceInference is running a trained model on an input to produce a prediction, generation, score, or other output.DatasetA dataset is an organized collection of examples or records used for training, evaluation, analysis, or system operation.EvaluationEvaluation is the structured measurement and investigation of how a model or system performs against defined questions, criteria, and contexts.

Compare training and inference

Sources and scope

What supports this object

Machine Learning Crash CourseGoogle for Developers

Explains training and evaluation in a practical technical course; specific procedures differ by model family and task.

Open source

Loading page content...

…