llama.cpp tools/quantize/quantize.cpp QUANT_OPTIONS table

Fields

id

sources/qlab2-quantize-options-table

URL

https://github.com/ggml-org/llama.cpp/blob/master/tools/quantize/quantize.cpp

Publisher

ggml-org

Kind

repo

Accessed

2026-10-03

notes

Canonical ftype->description table with the official per-type size and perplexity deltas vs fp16 on Llama-3-8B, and the official bpw strings for IQ/TQ types (IQ2_XXS 2.06, IQ2_XS 2.31, IQ2_S 2.5, IQ2_M 2.7, IQ1_S 1.56, IQ1_M 1.75, IQ3_XXS 3.06, IQ3_S 3.44, IQ3_M 3.66, IQ3_XS 3.3, IQ4_NL 4.50, IQ4_XS 4.25, TQ1_0 1.69, TQ2_0 2.06). | SOURCE AUDIT 2026-10-03: the 'published' field held a verbatim copy of this notes paragraph instead of a date (field corruption across the qlab2-* batch); published set to null because a GitHub README, doc page or source header has no reliable publication date of its own. Content of the document itself was re-verified live 2026-10-03 and the notes are correct.

Notes

Canonical ftype->description table with the official per-type size and perplexity deltas vs fp16 on Llama-3-8B, and the official bpw strings for IQ/TQ types (IQ2_XXS 2.06, IQ2_XS 2.31, IQ2_S 2.5, IQ2_M 2.7, IQ1_S 1.56, IQ1_M 1.75, IQ3_XXS 3.06, IQ3_S 3.44, IQ3_M 3.66, IQ3_XS 3.3, IQ4_NL 4.50, IQ4_XS 4.25, TQ1_0 1.69, TQ2_0 2.06). | SOURCE AUDIT 2026-10-03: the 'published' field held a verbatim copy of this notes paragraph instead of a date (field corruption across the qlab2-* batch); published set to null because a GitHub README, doc page or source header has no reliable publication date of its own. Content of the document itself was re-verified live 2026-10-03 and the notes are correct.

Cross-references

This record references no other records.