highperfocused/ollama
1
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2025-11-11 13:57:18 +01:00
Code Issues Packages Projects Releases Wiki Activity

ggml: Seperate tensor load from backend creation

Browse Source 
Currently, when the backend is created, the tensors are loaded at the
same time, which is a slow operation. This separates them to be two
steps:
 - Create backend, including enumerating tensors and memory allocation
 - Loading tensor data

This allows more flexibility in managing model loading.
This commit is contained in:
Jesse Gross
2025-04-17 13:42:40 -07:00
committed by Jesse Gross
parent d755577473
commit 94ab428e3f
13 changed files with 131 additions and 115 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
server/quantization_test.go
View File

@@ -271,7 +271,7 @@ func TestQuantizeModel(t *testing.T) {
t.Fatal(err.Error())
}
defer fp.Close()
meta, _, err := fsggml.Decode(fp, -1)
meta, err := fsggml.Decode(fp, -1)
if err != nil {
t.Fatal(err.Error())
}
@@ -303,7 +303,7 @@ func TestQuantizeModel(t *testing.T) {
t.Fatalf("failed to load the quantized model %s: %s", tmp.Name(), err)
}
defer fpNew.Close()
newMeta, _, err := fsggml.Decode(fpNew, -1)
newMeta, err := fsggml.Decode(fpNew, -1)
if err != nil {
t.Fatalf("failed to load the quantized model %s: %s", tmp.Name(), err)
}