Glossary Library

Technical GlossaryGenerative AI and LLM

INT4 Quantization

An aggressive quantization approach that reduces the model to 4-bit precision for much lower memory cost.

INT4 quantization is especially important for running large models on smaller hardware. It dramatically reduces memory cost, but it also carries a stronger risk of quality loss depending on task sensitivity. For that reason, calibration and careful benchmarking become especially critical at lower bit widths.

You Might Also Like

Explore these concepts to continue your artificial intelligence journey.

Glossary Cover

Checkpointed Backpropagation

A training technique that reduces memory usage by not storing all intermediate activations and recomputing them when needed.

Glossary Cover

uretken-yapay-zeka-ve-llm

Generative Model

A family of models that can generate new samples rather than only predicting labels.

Glossary Cover

uretken-yapay-zeka-ve-llm

Sampling

The process of making probabilistic choices from a learned distribution while generating new output.