try a fix

app.py

@@ -17,16 +17,19 @@ learn_plain = load_learner('xresnet50_pitch3.pkl')
 
 labels = learn_removeSilence.dls.vocab
 
-def process(Record, Upload, version):
-    if version == 'remove silence':
-        return predict(Record, Upload, learn_removeSilence)
-    elif version == 'plain':
-        return predict(Record, Upload, learn_plain)
+#def process(Record, Upload, version):
+ #   if version == 'remove silence':
+  #      return predict(Record, Upload, learn_removeSilence)
+   # elif version == 'plain':
+    #    return predict(Record, Upload, learn_plain)
 
-def predict(Record, Upload, learn):
+def predict(Record, Upload, version):
     if Upload: path = Upload
     else: path = Record
-    spec,pred,pred_idx,probs = learn.predict(str(path), with_input=True)
+    if version == 'remove silence':
+        spec,pred,pred_idx,probs = learn_removeSilence.predict(str(path), with_input=True)
+    elif version == 'plain':
+        spec,pred,pred_idx,probs = learn_plain.predict(str(path), with_input=True)
     fig,ax = plt.subplots(figsize=(16,10))
     show_image(spec, ax=ax)
     ax.invert_yaxis()
@@ -43,7 +46,7 @@ examples = [['代わる.mp3'],['大丈夫な.mp3'],['熱くない.mp3'], ['あ
 
 enable_queue=True
 
-gr.Interface(fn=process,
+gr.Interface(fn=predict,
 inputs=[gr.inputs.Audio(source='microphone', type='filepath', optional=True),
 gr.inputs.Audio(source='upload', type='filepath', optional=True),
 gr.inputs.Radio(choices=['plain','remove silence'], type="value", default='remove silence', label='version')