Upload 51 files

src/components/MultiSourceCaptioningView.tsx

@@ -505,7 +505,7 @@ export default function MultiSourceCaptioningView() {
       const output = results[session.outputNames[0]];
       const detections = postprocessYoloOutput(output);
       lastYoloBoxesRef.current = detections;
-      // Run FastVLM on the full frame (not boxes)
+      // Run FastVLM on the full frame (wait for YOLOv8 to finish)
       await runInference(processingVideo, prompt, (output: string) => {
         setDebugOutput(output);
       });