updates

utils/datasets.py

@@ -23,6 +23,7 @@ class ImageFolder():  # for eval-only
         self.nB = math.ceil(self.nF / batch_size)  # number of batches
         self.batch_size = batch_size
         self.height = img_size
+
         assert self.nF > 0, 'No images found in path %s' % path
 
         # RGB normalization values
@@ -65,7 +66,7 @@ class ListDataset():  # for training
         with open(path, 'r') as file:
             self.img_files = file.readlines()
 
-        if platform == 'darwin':  # macos
+        if platform == 'darwin':  # MacOS (local)
             self.img_files = [path.replace('\n', '').replace('/images', '/Users/glennjocher/Downloads/DATA/coco/images')
                               for path in self.img_files]
         else:  # linux (gcp cloud)
@@ -77,10 +78,10 @@ class ListDataset():  # for training
         self.nF = len(self.img_files)  # number of image files
         self.nB = math.ceil(self.nF / batch_size)  # number of batches
         self.batch_size = batch_size
-
-        # assert self.nB > 0, 'No images found in path %s' % path
         self.height = img_size
 
+        assert self.nB > 0, 'No images found in path %s' % path
+
         # RGB normalization values
         # self.rgb_mean = np.array([60.134, 49.697, 40.746], dtype=np.float32).reshape((1, 3, 1, 1))
         # self.rgb_std = np.array([29.99, 24.498, 22.046], dtype=np.float32).reshape((1, 3, 1, 1))

utils/utils.py

@@ -62,7 +62,7 @@ def weights_init_normal(m):
 
 
 def xyxy2xywh(x):  # Convert bounding box format from [x1, y1, x2, y2] to [x, y, w, h]
-    y = np.zeros(x.shape)
+    y = torch.zeros(x.shape) if x.dtype is torch.float32 else np.zeros(x.shape)
     y[:, 0] = (x[:, 0] + x[:, 2]) / 2
     y[:, 1] = (x[:, 1] + x[:, 3]) / 2
     y[:, 2] = x[:, 2] - x[:, 0]
@@ -71,11 +71,11 @@ def xyxy2xywh(x):  # Convert bounding box format from [x1, y1, x2, y2] to [x, y,
 
 
 def xywh2xyxy(x):  # Convert bounding box format from [x, y, w, h] to [x1, y1, x2, y2]
-    y = np.zeros(x.shape)
-    y[:, 0] = (x[:, 1] - x[:, 3] / 2)
-    y[:, 1] = (x[:, 2] - x[:, 4] / 2)
-    y[:, 2] = (x[:, 1] + x[:, 3] / 2)
-    y[:, 3] = (x[:, 2] + x[:, 4] / 2)
+    y = torch.zeros(x.shape) if x.dtype is torch.float32 else np.zeros(x.shape)
+    y[:, 0] = (x[:, 0] - x[:, 2] / 2)
+    y[:, 1] = (x[:, 1] - x[:, 3] / 2)
+    y[:, 2] = (x[:, 0] + x[:, 2] / 2)
+    y[:, 3] = (x[:, 1] + x[:, 3] / 2)
     return y