add more examples

app.py

@@ -28,7 +28,7 @@ state_dict = torch.load(ckpts[0], map_location=torch.device(device))['state_dict
 net = Model_LA(args, len(token_to_ix), pretrained_emb).to(device)
 net.load_state_dict(state_dict)
 
-def inference(video_path, text):
+def inference(source_video, transcription):
     # data preprocessing
     # text
     def clean(w):
@@ -38,10 +38,10 @@ def inference(video_path, text):
                 w.lower()
                 ).replace('-', ' ').replace('/', ' ')
 
-    s = [clean(w) for w in text.split() if clean(w) != '']
+    s = [clean(w) for w in transcription.split() if clean(w) != '']
 
     # Sound
-    _, mel, mag = load_spectrograms(video_path)
+    _, mel, mag = load_spectrograms(source_video)
 
     l_max_len = args.lang_seq_len
     a_max_len = args.audio_seq_len
@@ -70,6 +70,8 @@ title="Emotion Recognition"
 description=""
             
 examples = [              
+                ['examples/0h-zjBukYpk_2.mp4', "NOW IM NOT EVEN GONNA SUGAR COAT THIS THIS MOVIE FRUSTRATED ME TO SUCH AN EXTREME EXTENT THAT I WAS LOUDLY EXCLAIMING WHY AT THE END OF THE FILM"],
+                ['examples/0h-zjBukYpk_19.mp4', "NOW OTHER PERFORMANCES ARE BORDERLINE"],
                 ['examples/03bSnISJMiM_1.mp4', "IT WAS REALLY GOOD "],
                 ['examples/03bSnISJMiM_5.mp4', "AND THEY SHOULDVE I GUESS "],
            ]