Update app.py

app.py

@@ -8,13 +8,14 @@ import gradio as gr
 from mediapipe import Image as MPImage
 from mediapipe.tasks import python
 from mediapipe.tasks.python import vision
+import traceback
 
 # -----------------------------
 # 1. Paths & URLs
 # -----------------------------
 HAND_MODEL_PATH = "hand_landmarker.task"
 HAND_MODEL_URL = "https://storage.googleapis.com/mediapipe-models/hand_landmarker/hand_landmarker/float16/1/hand_landmarker.task"
-YOLO_MODEL_PATH = "yolov11n_finetuned_ASL.pt"  # Already in repo via Git LFS or small enough
+YOLO_MODEL_PATH = "yolov11n_finetuned_ASL.pt"
 
 # -----------------------------
 # 2. Download MediaPipe model if missing
@@ -39,44 +40,51 @@ hand_options = vision.HandLandmarkerOptions(base_options=base_options, num_hands
 detector = vision.HandLandmarker.create_from_options(hand_options)
 
 # -----------------------------
-# 4. Inference function
+# 4. Inference function with robust error handling
 # -----------------------------
 def predict_asl(image):
-    """
-    Input: numpy array (H x W x 3) from Gradio
-    Output: annotated image, predicted class, confidence
-    """
-    img = image.copy()
-    h, w, _ = img.shape
+    try:
+        if image is None:
+            raise ValueError("No image provided")
+        
+        img = image.copy()
+        h, w, _ = img.shape
+        print(f"🔹 Uploaded image shape: {img.shape}, dtype: {img.dtype}")
 
-    # --- Annotate hand landmarks ---
-    mp_image = MPImage.create_from_array(img)
-    detection_result = detector.detect(mp_image)
-    if detection_result.hand_landmarks:
-        for hand_landmarks in detection_result.hand_landmarks:
-            for landmark in hand_landmarks:
-                x, y = int(landmark.x * w), int(landmark.y * h)
-                cv2.circle(img, (x, y), 3, (0, 255, 0), -1)
+        # --- Annotate hand landmarks ---
+        mp_image = MPImage.create_from_array(img)
+        detection_result = detector.detect(mp_image)
+        if detection_result.hand_landmarks:
+            for hand_landmarks in detection_result.hand_landmarks:
+                for landmark in hand_landmarks:
+                    x, y = int(landmark.x * w), int(landmark.y * h)
+                    cv2.circle(img, (x, y), 3, (0, 255, 0), -1)
 
-    # --- YOLO prediction ---
-    results = yolo_model.predict(img, imgsz=300, verbose=False)[0]
-    pred_idx = results.probs.top1
-    pred_label = results.names[pred_idx]
-    confidence = results.probs.data[pred_idx].item()
+        # --- YOLO prediction ---
+        results = yolo_model.predict(img, imgsz=300, verbose=False)[0]
+        pred_idx = results.probs.top1
+        pred_label = results.names[pred_idx]
+        confidence = results.probs.data[pred_idx].item()
 
-    # Overlay prediction text
-    cv2.putText(
-        img,
-        f"{pred_label} ({confidence:.2f})",
-        (10, 30),
-        cv2.FONT_HERSHEY_SIMPLEX,
-        1,
-        (0, 0, 255),
-        2,
-        cv2.LINE_AA
-    )
+        # Overlay prediction text
+        cv2.putText(
+            img,
+            f"{pred_label} ({confidence:.2f})",
+            (10, 30),
+            cv2.FONT_HERSHEY_SIMPLEX,
+            1,
+            (0, 0, 255),
+            2,
+            cv2.LINE_AA
+        )
 
-    return cv2.cvtColor(img, cv2.COLOR_BGR2RGB), pred_label, round(confidence, 2)
+        return cv2.cvtColor(img, cv2.COLOR_BGR2RGB), pred_label, round(confidence, 2)
+
+    except Exception as e:
+        print("❌ Error in predict_asl:", e)
+        traceback.print_exc()
+        # Return original image and error placeholders
+        return image, "Error", 0.0
 
 # -----------------------------
 # 5. Gradio Interface