starred/ollama
1
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-04-27 19:25:55 +02:00
Code Issues Packages Projects Releases Wiki Activity

mlx: int4 groupsize 64 (#14682)

Browse Source 
Change affine 4bit integers to use groupsize 64
This commit is contained in:
Patrick Devine
2026-03-06 16:39:47 -08:00
committed by GitHub
parent 288077c3a3
commit e790dc435b
2 changed files with 3 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
x/create/client/quantize.go
View File

@@ -21,7 +21,7 @@ var quantizeParams = map[string]struct {
bits int
mode string
}{
"int4": {32, 4, "affine"},
"int4": {64, 4, "affine"},
"nvfp4": {16, 4, "nvfp4"},
"int8": {64, 8, "affine"},
"mxfp8": {32, 8, "mxfp8"},