Upload app.py with huggingface_hub

app.py

@@ -4,7 +4,6 @@ FaceAge-DINOv3 — Gradio demo for HuggingFace Spaces.
 Face detection : YuNet (OpenCV built-in, ~350 KB model, no extra deps)
 Age/gender     : FaceAge-DINOv3 ONNX (CPU, ~1.2 GB)
 """
-import urllib.request
 import os
 import numpy as np
 import gradio as gr
@@ -83,14 +82,11 @@ def _predict_crop(face_rgb: np.ndarray) -> dict:
 
 
 # ---------------------------------------------------------------------------
-# YuNet face detector  (cv2.FaceDetectorYN, OpenCV ≥ 4.5.4)
+# YuNet face detector  (cv2.FaceDetectorYN, loaded from HuggingFace Hub)
 # ---------------------------------------------------------------------------
 
-_YUNET_URL  = (
-    "https://github.com/opencv/opencv_zoo/raw/main/models/"
-    "face_detection_yunet/face_detection_yunet_2023mar.onnx"
-)
-_YUNET_PATH = "/tmp/face_detection_yunet_2023mar.onnx"
+_YUNET_REPO = "opencv/face_detection_yunet"
+_YUNET_FILE = "face_detection_yunet_2023mar.onnx"
 _DETECTOR   = None
 
 
@@ -99,21 +95,20 @@ def _load_detector():
     if _DETECTOR is not None:
         return
 
-    # Download model if not cached
-    if not os.path.exists(_YUNET_PATH):
-        print(f"[YuNet] Downloading model …")
-        try:
-            urllib.request.urlretrieve(_YUNET_URL, _YUNET_PATH)
-            print(f"[YuNet] Saved to {_YUNET_PATH}")
-        except Exception as e:
-            print(f"[YuNet] Download failed: {e} — face detection disabled")
-            _DETECTOR = "unavailable"
-            return
-
+    from huggingface_hub import hf_hub_download
     import cv2
+
+    try:
+        yunet_path = hf_hub_download(repo_id=_YUNET_REPO, filename=_YUNET_FILE)
+        print(f"[YuNet] Model: {yunet_path}")
+    except Exception as e:
+        print(f"[YuNet] Download failed: {e} — face detection disabled")
+        _DETECTOR = "unavailable"
+        return
+
     try:
         _DETECTOR = cv2.FaceDetectorYN.create(
-            model           = _YUNET_PATH,
+            model           = yunet_path,
             config          = "",
             input_size      = (320, 320),
             score_threshold = 0.6,
@@ -127,21 +122,21 @@ def _load_detector():
 
 
 def _detect_faces(img_rgb: np.ndarray,
-                  min_face_px: int = 20) -> list[tuple[int, int, int, int]]:
+                  min_face_px: int = 20,
+                  margin: int = 15) -> list[tuple[int, int, int, int]]:
     """
-    Returns list of (x0, y0, x1, y1) sorted by area (largest first).
+    Returns list of (x0, y0, x1, y1) with margin padding, sorted by area desc.
     Falls back to empty list if YuNet is unavailable.
     """
     if _DETECTOR == "unavailable" or _DETECTOR is None:
         return []
 
     import cv2
-    h, w  = img_rgb.shape[:2]
+    h, w    = img_rgb.shape[:2]
     img_bgr = cv2.cvtColor(img_rgb, cv2.COLOR_RGB2BGR)
 
-    # YuNet requires input_size to match the image
     _DETECTOR.setInputSize((w, h))
-    _, faces = _DETECTOR.detect(img_bgr)   # faces: None or Nx15
+    _, faces = _DETECTOR.detect(img_bgr)   # None or Nx15: [x,y,w,h, ...]
 
     if faces is None:
         return []
@@ -149,8 +144,11 @@ def _detect_faces(img_rgb: np.ndarray,
     bboxes = []
     for face in faces:
         x, y, fw, fh = face[:4].astype(int)
-        x0, y0 = max(0, x),      max(0, y)
-        x1, y1 = min(w, x + fw), min(h, y + fh)
+        # Add margin, clamp to image bounds
+        x0 = max(0, x - margin)
+        y0 = max(0, y - margin)
+        x1 = min(w, x + fw + margin)
+        y1 = min(h, y + fh + margin)
         if (x1 - x0) >= min_face_px and (y1 - y0) >= min_face_px:
             bboxes.append((x0, y0, x1, y1))