Update app.py

app.py

@@ -10,12 +10,10 @@ emotion_labels = {0: 'neutral', 1: 'calm', 2: 'happy', 3: 'sad', 4: 'angry', 5:
 def process_video_audio(video_path, audio_path):
 
     wav = pt.tensor(list(audio_path[1]))
-
-    print(wav.shape)
     
-    train_visual = np.zeros([1, 120, 120, 3, 10])
-    train_audio_wave = np.zeros([1, 261540])
-    train_audio_cnn = np.zeros([1, 150, 512, 1])
+    train_visual = np.zeros([5, 120, 120, 3, 10])
+    train_audio_wave = np.zeros([5, 261540])
+    train_audio_cnn = np.zeros([5, 150, 512, 1])
     
     mfcc = torchaudio.transforms.MFCC(n_mfcc=150, melkwargs={"n_fft": 1022, "n_mels": 150})
     
@@ -25,10 +23,11 @@ def process_video_audio(video_path, audio_path):
         train_audio_wave[0, :] = wav[:261540]
     else:
         train_audio_wave[0, :len(wav)] = wav[:]
-    print()
+    
+    print(train_audio_wave.shape)
     print(type(mfcc(train_audio_wave)))
     
-    train_audio_cnn[0, :, :, 0] = mfcc(train_audio_wave)
+    train_audio_cnn[:, :, :, 0] = mfcc(train_audio_wave)
 
     print(train_audio_cnn[0].shape)