feat: align with the model-spec v0.0.2 (#87)

Signed-off-by: chlins <[email protected]>

Author: chlins

go.mod

@@ -3,7 +3,7 @@ module github.com/CloudNativeAI/modctl
 go 1.23.3
 
 require (
-	github.com/CloudNativeAI/model-spec v0.0.1
+	github.com/CloudNativeAI/model-spec v0.0.2
 	github.com/distribution/distribution/v3 v3.0.0-rc.3
 	github.com/distribution/reference v0.6.0
 	github.com/dustin/go-humanize v1.0.1

go.sum

@@ -1,5 +1,5 @@
-github.com/CloudNativeAI/model-spec v0.0.1 h1:BgVIStKTLuL1DrLC5A/gmHcR8TEhFCDz9+fYdCUa/CY=
-github.com/CloudNativeAI/model-spec v0.0.1/go.mod h1:3U/4zubBfbUkW59ATSg41HnkYyKrKUcKFH/cVdoPQnk=
+github.com/CloudNativeAI/model-spec v0.0.2 h1:uCO86kMk8wwadn8vKs0wT4petig5crByTIngdO3L2cQ=
+github.com/CloudNativeAI/model-spec v0.0.2/go.mod h1:3U/4zubBfbUkW59ATSg41HnkYyKrKUcKFH/cVdoPQnk=
 github.com/VividCortex/ewma v1.2.0 h1:f58SaIzcDXrSy3kWaHNvuJgJ3Nmz59Zji6XoJR/q1ow=
 github.com/VividCortex/ewma v1.2.0/go.mod h1:nz4BbCtbLyFDeC9SUHbtcT5644juEuWfUAUnGx7j5l4=
 github.com/acarl005/stripansi v0.0.0-20180116102854-5a71ef0e047d h1:licZJFw2RwpHMqeKTCYkitsPqHNxTmd4SNR5r94FGM8=

pkg/backend/build/build.go

@@ -22,7 +22,6 @@ import (
 	"encoding/json"
 	"fmt"
 	"path/filepath"
-	"strconv"
 	"time"
 
 	"github.com/CloudNativeAI/modctl/pkg/archiver"
@@ -74,13 +73,8 @@ func buildModelConfig(modelfile modelfile.Modelfile) (*modelspec.Model, error) {
 		Format:       modelfile.GetFormat(),
 		Precision:    modelfile.GetPrecision(),
 		Quantization: modelfile.GetQuantization(),
+		ParamSize:    modelfile.GetParamsize(),
 	}
-	// parse the parameter size.
-	paramSize, err := strconv.ParseUint(modelfile.GetParamsize(), 10, 64)
-	if err != nil {
-		return nil, fmt.Errorf("failed to parse paramsize %s to uint64: %w", modelfile.GetParamsize(), err)
-	}
-	config.ParameterSize = paramSize
 
 	createdAt := time.Now()
 	descriptor := modelspec.ModelDescriptor{

pkg/backend/inspect.go

@@ -99,7 +99,7 @@ func (b *backend) Inspect(ctx context.Context, target string) (*InspectedModelAr
 		Family:       config.Descriptor.Family,
 		Format:       config.Config.Format,
 		Name:         config.Descriptor.Name,
-		ParamSize:    fmt.Sprintf("%d", config.Config.ParameterSize),
+		ParamSize:    config.Config.ParamSize,
 		Precision:    config.Config.Precision,
 		Quantization: config.Config.Quantization,
 	}

pkg/backend/inspect_test.go

@@ -120,7 +120,7 @@ func TestInspect(t *testing.T) {
   "config": {
     "architecture": "transformer",
     "format": "tensorflow",
-    "parameterSize": 50000000000,
+    "paramSize": "0.5b",
     "precision": "int8",
     "puantization": "gptq"
   }
@@ -138,7 +138,7 @@ func TestInspect(t *testing.T) {
 	assert.Equal(t, "qwen2", inspected.Family)
 	assert.Equal(t, "tensorflow", inspected.Format)
 	assert.Equal(t, "Qwen2.5-0.5B", inspected.Name)
-	assert.Equal(t, "50000000000", inspected.ParamSize)
+	assert.Equal(t, "0.5b", inspected.ParamSize)
 	assert.Equal(t, "int8", inspected.Precision)
 	assert.Equal(t, "gptq", inspected.Quantization)
 	assert.Len(t, inspected.Layers, 8)