Update app.py

app.py

@@ -17,6 +17,18 @@ LABEL_MAPS = {
     'UCF' : UCF_label_map 
 }
 
+ALL_MODELS = [
+    'TFVideoMAE_L_K400_16x224',
+    'TFVideoMAE_B_SSv2_16x224',
+    'TFVideoMAE_B_UCF_16x224',
+]
+
+sample_example = [
+    ["examples/k400.mp4",  ALL_MODELS[0], 0.9],
+    ["examples/ssv2.avi",  ALL_MODELS[1], 0.8],
+    ["examples/ucf.mp4",   ALL_MODELS[2], 0.7],
+]
+
 def tube_mask_generator(mask_ratio):
     window_size = (
         num_frames // 2, 
@@ -58,10 +70,6 @@ def get_model(model_type):
 
 
 def inference(video_file, model_type, mask_ratio):
-    print(video_file)
-    print(model_type)
-    print(mask_ratio)
-    print('---------------')
     # get sample data
     container = read_video(video_file)
     frames = frame_sampling(container, num_frames=num_frames)
@@ -101,20 +109,6 @@ def inference(video_file, model_type, mask_ratio):
 
 
 def main():
-    datasets = ['K400', 'SSv2', 'UCF']
-
-    ALL_MODELS = [
-        'TFVideoMAE_L_K400_16x224',
-        'TFVideoMAE_B_SSv2_16x224',
-        'TFVideoMAE_B_UCF_16x224',
-    ]
-
-    sample_example = [
-        ["examples/k400.mp4",  ALL_MODELS[0], 0.9],
-        ["examples/ssv2.webm",  ALL_MODELS[1], 0.8],
-        ["examples/ucf.mp4",  ALL_MODELS[2], 0.7],
-    ]
-
     iface = gr.Interface(
         fn=inference,
         inputs=[ 
@@ -124,12 +118,6 @@ def main():
                 value="TFVideoMAE_S_K400_16x224", 
                 label="Model"
             ),
-            # gr.Radio(
-            #     datasets, 
-            #     type='value',
-            #     default=datasets[0],
-            #     label='Dataset', 
-            # ),
             gr.Slider(
                 0.5,
                 1.0,