llama.cpp
convert : handle compressed-tensors quant method
#17069

Merged

convert : handle compressed-tensors quant method #17069

compilade merged 6 commits into master from compilade/convert-prequant-compressed-tensors

convert : handle compressed-tensors quant method

33dba6ce

convert : handle int-quantized models

d23bdd57

convert : handle naive-quantized models

33dcb44a

gguf-py : __pos__ is also unary

987862ad

convert : fix flake8 lint

3770d941

convert : use F32 for dequant of pack-quantized tensors

128118fd

compilade requested a review from

CISC 102 days ago

github-actions added python

compilade added enhancement

ggerganov requested a review from

ngxson 101 days ago

CISC approved these changes on 2025-11-07

ngxson approved these changes on 2025-11-07

compilade merged 1c07c0c6 into master 99 days ago

Reviewers

ngxson

CISC

Assignees

No one assigned

Labels

enhancement python

Milestone

No milestone