Commit e835ef18 authored by Michael Yang's avatar Michael Yang
Browse files

fix: quantization with template

parent c7c2f3bc
...@@ -414,17 +414,22 @@ func CreateModel(ctx context.Context, name model.Name, modelFileDir, quantizatio ...@@ -414,17 +414,22 @@ func CreateModel(ctx context.Context, name model.Name, modelFileDir, quantizatio
return err return err
} }
layers, err := parseFromFile(ctx, temp, "", fn) layer, err := NewLayer(temp, baseLayer.MediaType)
if err != nil { if err != nil {
return err return err
} }
if len(layers) != 1 { if _, err := temp.Seek(0, io.SeekStart); err != nil {
return errors.New("quantization failed") return err
}
ggml, _, err := llm.DecodeGGML(temp)
if err != nil {
return err
} }
baseLayer.Layer = layers[0].Layer baseLayer.Layer = layer
baseLayer.GGML = layers[0].GGML baseLayer.GGML = ggml
} }
} }
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment