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

enable q8, q5, 5_1, and f32 for linux gpu (#699)

Browse source
This commit is contained in:
Bruce MacDonald 2023-10-05 12:53:47 -04:00 committed by GitHub
parent d104b7e997
commit d06bc0cb6e
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
1 changed files with 17 additions and 14 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

3
llm/llm.go
View file

@ -5,6 +5,7 @@ import (
"fmt" "fmt"
"log" "log"
"os" "os"
"runtime"
"github.com/pbnjay/memory" "github.com/pbnjay/memory"
@ -37,6 +38,7 @@ func New(workDir, model string, adapters []string, opts api.Options) (LLM, error
return nil, err return nil, err
} }
if runtime.GOOS == "darwin" {
switch ggml.FileType() { switch ggml.FileType() {
case "Q8_0": case "Q8_0":
if ggml.Name() != "gguf" && opts.NumGPU != 0 { if ggml.Name() != "gguf" && opts.NumGPU != 0 {
@ -53,6 +55,7 @@ func New(workDir, model string, adapters []string, opts api.Options) (LLM, error
opts.NumGPU = 0 opts.NumGPU = 0
} }
} }
}
totalResidentMemory := memory.TotalMemory() totalResidentMemory := memory.TotalMemory()
switch ggml.ModelType() { switch ggml.ModelType() {