690618:2126 239 #01

backend/src/modules/ai/services/vram-monitor.service.ts

@@ -17,7 +17,11 @@ export interface VramStatus {
   totalVramMb: number;
   usedVramMb: number;
   freeVramMb: number;
-  loadedModels: string[];
+  loadedModels: Array<{
+    modelId: string;
+    modelName: string;
+    vramUsageMB: number;
+  }>;
   hasCapacity: boolean;
 }
 
@@ -102,7 +106,11 @@ export class VramMonitorService {
         }>;
       }>(`${this.ollamaUrl}/api/ps`, { timeout: 3000 });
       const models = response.data?.models ?? [];
-      const loadedModels = models.map((m) => m.name);
+      const loadedModels = models.map((m) => ({
+        modelId: m.name,
+        modelName: m.name,
+        vramUsageMB: Math.round((m.size_vram || 0) / (1024 * 1024)),
+      }));
       const headroom = await this.getVramHeadroom();
       return {
         totalVramMb: headroom.totalMb,

backend/src/modules/ai/tests/vram-monitor.service.spec.ts

@@ -120,7 +120,13 @@ describe('VramMonitorService', () => {
           },
         });
       const status = await service.getVramStatus(4000);
-      expect(status.loadedModels).toContain('np-dms-ai:latest');
+      expect(status.loadedModels).toEqual([
+        {
+          modelId: 'np-dms-ai:latest',
+          modelName: 'np-dms-ai:latest',
+          vramUsageMB: 3072,
+        },
+      ]);
       expect(status.totalVramMb).toBe(8192);
       expect(status.hasCapacity).toBe(true); // 8192MB - 3072MB = 5120MB free > 4000MB required
     });