Update app.py

app.py

@@ -31,18 +31,31 @@ def text2story(text):
 # text2audio
 def text2audio(story_text):
     
-    tts_pipeline = pipeline("text-to-speech", model="suno/bark-small")
+    # tts_pipeline = pipeline("text-to-speech", model="suno/bark-small")
 
-    audio_data = tts_pipeline(story_text)
+    # audio_data = tts_pipeline(story_text)
 
+    # audio_buffer = io.BytesIO()
+    # wavfile.write(audio_buffer, rate=audio_data["sampling_rate"], data=audio_data["audio"])
+    # audio_buffer.seek(0)
+
+    # return {
+    #     'audio': audio_buffer.getvalue(),  
+    #     'sampling_rate': audio_data["sampling_rate"] 
+    # }
+
+
+    inputs = processor(text=story_text, return_tensors="pt")
+    with torch.no_grad():
+        speech = model.generate(**inputs)
+
+    audio_data = speech.cpu().numpy().squeeze()
+    
     audio_buffer = io.BytesIO()
-    wavfile.write(audio_buffer, rate=audio_data["sampling_rate"], data=audio_data["audio"])
+    wavfile.write(audio_buffer, rate=16000, data=audio_data)  # 16kHz 采样率
     audio_buffer.seek(0)
 
-    return {
-        'audio': audio_buffer.getvalue(),  
-        'sampling_rate': audio_data["sampling_rate"] 
-    }
+    return {'audio': audio_buffer.getvalue(), 'sampling_rate': 16000}
 
 
 # program main part