Skip to content

add modelcar image size & byte size #2808

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
adysenrothman wants to merge 1 commit into
base: main
Choose a base branch
from
+74 −23
Open

add modelcar image size & byte size #2808

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
33 changes: 26 additions & 7 deletions catalog/internal/catalog/modelcatalog/performance_metrics.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -22,13 +22,15 @@ import (
// metadataJSON represents the minimal structure needed from metadata.json files
// Only the ID field is needed to look up existing models
type metadataJSON struct {
ID string `json:"id"` // Maps to model name for lookup
OverallAccuracy *float64 `json:"overall_accuracy"` // Overall accuracy score for the model
Size *string `json:"size"` // Model parameter count (e.g., "8B params")
TensorType *string `json:"tensor_type"` // Data precision (e.g., "FP16", "INT4")
VariantGroupID *string `json:"variant_group_id"` // UUID linking model variants together
MinVRAMGB *string `json:"min_vram_gb"` // Minimum VRAM required (e.g., "265 GB")
ColdStartMatrix []coldStartEntry `json:"cold_start_matrix"` // Cold start times per GPU configuration
ID string `json:"id"` // Maps to model name for lookup
OverallAccuracy *float64 `json:"overall_accuracy"` // Overall accuracy score for the model
Size *string `json:"size"` // Model parameter count (e.g., "8B params")
TensorType *string `json:"tensor_type"` // Data precision (e.g., "FP16", "INT4")
VariantGroupID *string `json:"variant_group_id"` // UUID linking model variants together
MinVRAMGB *string `json:"min_vram_gb"` // Minimum VRAM required (e.g., "265 GB")
ModelcarImageSize *string `json:"modelcar_image_size"` // Human-readable modelcar image size (e.g., "230.17 GB")
ModelcarImageSizeBytes *int64 `json:"modelcar_image_size_bytes"` // Modelcar image size in bytes
ColdStartMatrix []coldStartEntry `json:"cold_start_matrix"` // Cold start times per GPU configuration
}

type coldStartEntry struct {
Expand Down Expand Up @@ -765,6 +767,23 @@ func enrichCatalogModelFromMetadata(existingModel dbmodels.CatalogModel, metadat
})
}

if metadata.ModelcarImageSize != nil && *metadata.ModelcarImageSize != "" {
customProperties = append(customProperties, models.Properties{
Name: "modelcar_image_size",
StringValue: metadata.ModelcarImageSize,
IsCustomProperty: true,
})
}

if metadata.ModelcarImageSizeBytes != nil {
bytesStr := strconv.FormatInt(*metadata.ModelcarImageSizeBytes, 10)
customProperties = append(customProperties, models.Properties{
Name: "modelcar_image_size_bytes",
StringValue: &bytesStr,
IsCustomProperty: true,
})
}

if len(customProperties) == 0 {
return nil // Nothing to update
}
Expand Down
64 changes: 48 additions & 16 deletions catalog/internal/catalog/modelcatalog/performance_metrics_test.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -956,15 +956,17 @@ func TestUnmarshalJSON_EdgeCases(t *testing.T) {

func TestParseMetadataJSON_NewFields(t *testing.T) {
tests := []struct {
name string
jsonData string
wantID string
wantSize *string
wantTensorType *string
wantVariantID *string
wantMinVRAMGB *string
wantColdStartMatrix []coldStartEntry
wantErr bool
name string
jsonData string
wantID string
wantSize *string
wantTensorType *string
wantVariantID *string
wantMinVRAMGB *string
wantModelcarImageSize *string
wantModelcarImageSizeBytes *int64
wantColdStartMatrix []coldStartEntry
wantErr bool
}{
{
name: "complete metadata with all new fields",
Expand Down Expand Up @@ -1124,6 +1126,8 @@ func TestParseMetadataJSON_NewFields(t *testing.T) {
"tensor_type": "FP8",
"variant_group_id": "vwx234mn-5678-901v-wx23-456789abcdef",
"min_vram_gb": "265 GB",
"modelcar_image_size": "230.17 GB",
"modelcar_image_size_bytes": 230171650363,
"cold_start_matrix": [
{
"gpu_type": "A100-80",
Expand All @@ -1137,11 +1141,13 @@ func TestParseMetadataJSON_NewFields(t *testing.T) {
}
]
}`,
wantID: "sample-model/test-405b-instruct",
wantSize: &[]string{"405B params"}[0],
wantTensorType: &[]string{"FP8"}[0],
wantVariantID: &[]string{"vwx234mn-5678-901v-wx23-456789abcdef"}[0],
wantMinVRAMGB: &[]string{"265 GB"}[0],
wantID: "sample-model/test-405b-instruct",
wantSize: &[]string{"405B params"}[0],
wantTensorType: &[]string{"FP8"}[0],
wantVariantID: &[]string{"vwx234mn-5678-901v-wx23-456789abcdef"}[0],
wantMinVRAMGB: &[]string{"265 GB"}[0],
wantModelcarImageSize: &[]string{"230.17 GB"}[0],
wantModelcarImageSizeBytes: &[]int64{230171650363}[0],
wantColdStartMatrix: []coldStartEntry{
{GPUType: "A100-80", GPUCount: "4", ColdStartTimeToLoadSeconds: "587.3"},
{GPUType: "B200", GPUCount: "2", ColdStartTimeToLoadSeconds: "559.9"},
Expand Down Expand Up @@ -1220,6 +1226,16 @@ func TestParseMetadataJSON_NewFields(t *testing.T) {
t.Errorf("parseMetadataJSON() MinVRAMGB = %v, want %v", got.MinVRAMGB, tt.wantMinVRAMGB)
}

// Test ModelcarImageSize field
if (got.ModelcarImageSize == nil) != (tt.wantModelcarImageSize == nil) || (got.ModelcarImageSize != nil && tt.wantModelcarImageSize != nil && *got.ModelcarImageSize != *tt.wantModelcarImageSize) {
t.Errorf("parseMetadataJSON() ModelcarImageSize = %v, want %v", got.ModelcarImageSize, tt.wantModelcarImageSize)
}

// Test ModelcarImageSizeBytes field
if (got.ModelcarImageSizeBytes == nil) != (tt.wantModelcarImageSizeBytes == nil) || (got.ModelcarImageSizeBytes != nil && tt.wantModelcarImageSizeBytes != nil && *got.ModelcarImageSizeBytes != *tt.wantModelcarImageSizeBytes) {
t.Errorf("parseMetadataJSON() ModelcarImageSizeBytes = %v, want %v", got.ModelcarImageSizeBytes, tt.wantModelcarImageSizeBytes)
}

// Test ColdStartMatrix field
if len(got.ColdStartMatrix) != len(tt.wantColdStartMatrix) {
t.Errorf("parseMetadataJSON() ColdStartMatrix length = %d, want %d", len(got.ColdStartMatrix), len(tt.wantColdStartMatrix))
Expand Down Expand Up @@ -1477,9 +1493,13 @@ func TestEnrichCatalogModelFromMetadata_NewFields(t *testing.T) {
}

minVRAM := "80GB"
modelcarSize := "230.17 GB"
modelcarSizeBytes := int64(230171650363)
metadata := metadataJSON{
ID: modelName,
MinVRAMGB: &minVRAM,
ID: modelName,
MinVRAMGB: &minVRAM,
ModelcarImageSize: &modelcarSize,
ModelcarImageSizeBytes: &modelcarSizeBytes,
ColdStartMatrix: []coldStartEntry{
{GPUType: "A100", GPUCount: "2", ColdStartTimeToLoadSeconds: "127.3"},
{GPUType: "H100", GPUCount: "1", ColdStartTimeToLoadSeconds: "68.9"},
Expand Down Expand Up @@ -1510,6 +1530,18 @@ func TestEnrichCatalogModelFromMetadata_NewFields(t *testing.T) {
} else if v != "80GB" {
t.Errorf("min_vram_gb = %q, want %q", v, "80GB")
}

if v, ok := propMap["modelcar_image_size"]; !ok {
t.Error("expected custom property 'modelcar_image_size' to be set")
} else if v != "230.17 GB" {
t.Errorf("modelcar_image_size = %q, want %q", v, "230.17 GB")
}

if v, ok := propMap["modelcar_image_size_bytes"]; !ok {
t.Error("expected custom property 'modelcar_image_size_bytes' to be set")
} else if v != "230171650363" {
t.Errorf("modelcar_image_size_bytes = %q, want %q", v, "230171650363")
}
}

func TestCreateColdStartArtifact(t *testing.T) {
Expand Down
Loading