123456

app.py

@@ -62,6 +62,9 @@ vocoder.load_state_dict(torch.load(HIFIGAN_CHECKPT, map_location=lambda loc, sto
 _ = vocoder.to(device).eval()
 vocoder.remove_weight_norm()
 
+emotions = sorted(["angry", "surprise", "fear", "happy", "neutral", "sad"]) 
+spekears = ['Madi', 'Marzhan', 'Akzhol']
+
 def generate_audio(text, quantity, speaker, emotion_1, emotion_2):
     x, x_lengths = convert_text(text)
     emo_1, emo_2 = emotion_1, emotion_2
@@ -92,14 +95,11 @@ def generate_audio(text, quantity, speaker, emotion_1, emotion_2):
     audio = audio * 32768.0
     audio = audio.detach().cpu().numpy().astype('int16')
     sr = 22050
-    return sr, audio
+    return (sr, audio)
 
 # def sentence_builder(quantity, emotion_1, emotion_2):
 #     return f"""The {quantity} {emotion_1}s from {" and ".join(emotion_2)}"""
 
-emotions = sorted(["angry", "surprise", "fear", "happy", "neutral", "sad"]) 
-spekears = ['Madi', 'Marzhan', 'Akzhol']
-
 demo = gr.Interface(
     generate_audio,
     [
@@ -111,7 +111,7 @@ demo = gr.Interface(
         gr.Dropdown(emotions, value=emotions[3], label="Emotion 2", info="Select second emotion."
         ),
     ],
-    "audio"
+    "audio",
 )
 
 demo.launch()