llama.cpp
IQ4_NL: 4-bit non-linear quants with blocks of 32
#5590
Merged

IQ4_NL: 4-bit non-linear quants with blocks of 32 #5590

ikawrakow merged 6 commits into master from ik/iq4_nl_no_superblock
ikawrakow
iq4_nl: squash commits for easier rebase
9b0d3a85
iq4_nl: Fix after merging with master
1d900212
iq4_nl: another fix after merging with master
e7b999c3
Use IQ4_NL instead of Q4_K when using k-quants is not possible
3fc45558
sorasoras
ikawrakow
sorasoras
Artefact2
Fix typo that makes several tests fail
b376bbb2
It was the ggml_vdotq thing missed inside the brackets
daacf6ca
sorasoras
JianbangZ
ggerganov
ggerganov approved these changes on 2024-02-20
sorasoras
ikawrakow ikawrakow merged a14679cc into master 1 year ago
ikawrakow ikawrakow deleted the ik/iq4_nl_no_superblock branch 1 year ago
JianbangZ
EinhartStratos

Login to write a write a comment.

Login via GitHub

Reviewers
Assignees
No one assigned
Labels
Milestone