V — The Systems That Run Them → Chapter 26
FROM SYSTEMS TO FRONTIER ML

Quantization in practice

PTQ basics (int8/int4, scale + zero point), the LLM.int8/GPTQ/AWQ family, the GGML/llama.cpp quantization family (q4_0..q6_K, q4_K_M vs q4_K_S, IQ-quants, imatrix), and quantization-aware training (STE, BitNet, QLoRA).

§1 PTQ basics — int8, int4, scales, zero points §2 The GGML quantization family — q4_0 to q6_K, and the q4_K_M naming §3 Quantization-aware training — STE, BitNet, QLoRA

← ALL CHAPTERS