baalajimaestro/ollama
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

Merge pull request #4624 from ollama/mxyng/fix-5

Browse source 
fix q5_0, q5_1
This commit is contained in:
Michael Yang 2024-05-24 16:11:21 -07:00 committed by GitHub
commit 9a3c8003c8
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
1 changed files with 2 additions and 2 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

4
llm/ggml.go
View file

@ -125,9 +125,9 @@ type Tensor struct {
func (t Tensor) blockSize() uint64 {
switch t.Kind {
case 0, 1, 24, 25, 26, 27, 28, 31: // F32, F16, I8, I16, I32, I64, F64, BF16
case 0, 1, 24, 25, 26, 27, 28, 30: // F32, F16, I8, I16, I32, I64, F64, BF16
return 1
case 2, 3, 8, 9, 20: // Q4_0, Q4_1, Q8_0, Q8_1, IQ4_NL
case 2, 3, 4, 5, 6, 7, 8, 9, 20: // Q4_0, Q4_1, Q5_0, Q5_1, Q8_0, Q8_1, IQ4_NL
return 32
default: // All others
return 256