change net_input to 112 according to r(2+1)d, accelerate training

lib/dataset.py

@@ -12,8 +12,8 @@ def __init__(self, directory, mode='train', clip_len=8, frame_sample_rate=1):
         folder = Path(directory)/mode  # get the directory of the specified split
         self.clip_len = clip_len
 
-        self.short_side = [256, 320]
-        self.crop_size = 224
+        self.short_side = [128, 160]
+        self.crop_size = 112
         self.frame_sample_rate = frame_sample_rate
         self.mode = mode
 
@@ -28,7 +28,8 @@ def __init__(self, directory, mode='train', clip_len=8, frame_sample_rate=1):
         # convert the list of label names into an array of label indices
         self.label_array = np.array([self.label2index[label] for label in labels], dtype=int)
 
-        with open('labels.txt', 'w') as f:
+        label_file = str(os.listdir(folder))+'class_labels.txt'
+        with open(label_file, 'w') as f:
             for id, label in enumerate(sorted(self.label2index)):
                 f.writelines(str(id + 1) + ' ' + label + '\n')