English
← Back to AI Technology

Inference and deployment

Quantization

Representing weights, activations, or caches with lower numerical precision. It can reduce memory, bandwidth, and compute costs, with accuracy and hardware-support tradeoffs.

Reference period
2015 / LLM PTQ 2022
Last reviewed

Key terms

  • precision
  • INT8
  • INT4
  • weight-only

Connected across the project

Primary sources

  1. Deep Compression
  2. GPTQ
  3. Transformers quantization documentation

This is a curated technical map, not a claim of comprehensive coverage.