Adding link to Google drive for non Baidu users

README.md

@@ -2,7 +2,7 @@
 [A PyTorch Implementation of Dual Shot Face Detector](https://arxiv.org/abs/1810.10220?utm_source=feedburner&utm_medium=feed&utm_campaign=Feed%3A+arxiv%2FQSXk+%28ExcitingAds%21+cs+updates+on+arXiv.org%29)
 
 ### Description
-I use basenet [vgg](https://pan.baidu.com/s/1Q-YqoxJyqvln6KTcIck1tQ) to train DSFD,the model can be downloaded in [DSFD](https://pan.baidu.com/s/17cpDHEwYVxWmOIPqUy5zCQ).the AP in WIDER FACE as following:  
+I use basenet [vgg](https://pan.baidu.com/s/1Q-YqoxJyqvln6KTcIck1tQ) to train DSFD, the model can be downloaded in [DSFD with Baidu](https://pan.baidu.com/s/17cpDHEwYVxWmOIPqUy5zCQ) or [DSFD with Google Drive](https://drive.google.com/open?id=11pZy4DhslDP9cEk2uN9wNhTmB1VZ7GIR). The AP in WIDER FACE as following:  
 
 | Easy MAP | Medium MAP	|  hard MAP |
 | ---------|------------| --------- |
@@ -68,4 +68,4 @@ python demo.py
 
 ### References
 * [Dual Shot Face Detector](https://arxiv.org/abs/1810.10220?utm_source=feedburner&utm_medium=feed&utm_campaign=Feed%3A+arxiv%2FQSXk+%28ExcitingAds%21+cs+updates+on+arXiv.org%29)
-* [ssd.pytorch](https://github.com/amdegroot/ssd.pytorch)
+* [ssd.pytorch](https://github.com/amdegroot/ssd.pytorch)

demo.py

@@ -33,7 +33,7 @@
                     help='Directory for detect result')
 parser.add_argument('--model',
                     type=str,
-                    default='weights/dsfd_face.pth', help='trained model')
+                    default='weights/dsfd_vgg_0.880.pth', help='trained model')
 parser.add_argument('--thresh',
                     default=0.4, type=float,
                     help='Final confidence threshold')
@@ -58,10 +58,8 @@ def detect(net, img_path, thresh):
 
     img = np.array(img)
     height, width, _ = img.shape
-    max_im_shrink = np.sqrt(
-        1500 * 1000 / (img.shape[0] * img.shape[1]))
-    image = cv2.resize(img, None, None, fx=max_im_shrink,
-                       fy=max_im_shrink, interpolation=cv2.INTER_LINEAR)
+    max_im_shrink = np.sqrt(1500 * 1000 / (img.shape[0] * img.shape[1]))
+    image = cv2.resize(img, None, None, fx=max_im_shrink,fy=max_im_shrink, interpolation=cv2.INTER_LINEAR)
 
     x = to_chw_bgr(image)
     x = x.astype('float32')
@@ -88,31 +86,32 @@ def detect(net, img_path, thresh):
             j += 1
             cv2.rectangle(img, left_up, right_bottom, (0, 0, 255), 2)
             conf = "{:.2f}".format(score)
-            text_size, baseline = cv2.getTextSize(
-                conf, cv2.FONT_HERSHEY_SIMPLEX, 0.3, 1)
+            text_size, baseline = cv2.getTextSize(conf, cv2.FONT_HERSHEY_SIMPLEX, 0.3, 1)
             p1 = (left_up[0], left_up[1] - text_size[1])
-            cv2.rectangle(img, (p1[0] - 2 // 2, p1[1] - 2 - baseline),
-                          (p1[0] + text_size[0], p1[1] + text_size[1]),[255,0,0], -1)
-            cv2.putText(img, conf, (p1[0], p1[
-                            1] + baseline), cv2.FONT_HERSHEY_SIMPLEX, 0.3, (255, 255, 255), 1, 8)
+            cv2.rectangle(img, (p1[0] - 2 // 2, p1[1] - 2 - baseline),(p1[0] + text_size[0], p1[1] + text_size[1]),[255,0,0], -1)
+            cv2.putText(img, conf, (p1[0], p1[1] + baseline), cv2.FONT_HERSHEY_SIMPLEX, 0.3, (255, 255, 255), 1, 8)
 
     t2 = time.time()
     print('detect:{} timer:{}'.format(img_path, t2 - t1))
+    print('Found:{}'.format(j))
 
     cv2.imwrite(os.path.join(args.save_dir, os.path.basename(img_path)), img)
 
 
 if __name__ == '__main__':
     net = build_net('test', cfg.NUM_CLASSES, args.network)
-    net.load_state_dict(torch.load(args.model))
+    if use_cuda:
+        net.load_state_dict(torch.load(args.model))
+    else:
+        net.load_state_dict(torch.load(args.model, map_location=lambda storage, loc: storage))
+
     net.eval()
 
     if use_cuda:
         net.cuda()
         cudnn.benckmark = True
 
     img_path = './img'
-    img_list = [os.path.join(img_path, x)
-                for x in os.listdir(img_path) if x.endswith('jpg')]
+    img_list = [os.path.join(img_path, x) for x in os.listdir(img_path) if x.endswith('jpg')]
     for path in img_list:
         detect(net, path, args.thresh)

layers/functions/__init__.py

@@ -1,5 +1,5 @@
 from .prior_box import PriorBox 
-from detection import Detect
+from .detection import Detect
 
 __all__=['Detect','PriorBox']
 

weights/README.md

@@ -0,0 +1 @@
+Github does not allow files over 100MB. For this reason, please download the model on [Baidu](https://pan.baidu.com/s/17cpDHEwYVxWmOIPqUy5zCQ) or [Google](https://drive.google.com/open?id=11pZy4DhslDP9cEk2uN9wNhTmB1VZ7GIR) and place it here.