Cleanup

app.py

@@ -7,16 +7,13 @@ p = pipeline("audio-classification", model=model_name)
 
 def process(file):
         pred = p(file)
-        labels = dict()
-        for l in pred:
-            labels[l['label']]=l['score']
-        return labels
+        return { x['label']:x['score'] for x in pred}
 
 # Gradio inputs
-mic = gr.inputs.Audio(source='microphone', type='filepath', label='Speech input', optional=True)
+mic = gr.inputs.Audio(source='microphone', type='filepath', label='Speech input')
 
 # Gradio outputs
-keyword = gr.outputs.Label(num_top_classes=3)
+labels = gr.outputs.Label(num_top_classes=3)
 
 description = "This Space showcases a wav2vec2-conformer-rel-pos-large model fine-tuned for audio classification on the speech_commands dataset. \n \n It can spot one of the following keywords: 'Yes', 'No', 'Up', 'Down', 'Left', 'Right', 'On', 'Off', 'Stop', 'Go', 'Zero', 'One', 'Two', 'Three', 'Four', 'Five', 'Six', 'Seven', 'Eight', 'Nine', 'Bed', 'Bird', 'Cat', 'Dog', 'Happy', 'House', 'Marvin', 'Sheila', 'Tree', 'Wow', 'Backward', 'Forward', 'Follow', 'Learn', 'Visual'."
 
@@ -24,9 +21,8 @@ iface = gr.Interface(
 	theme='huggingface',
 	description=description,
 	fn=process,
-        layout='horizontal',
 	inputs=[mic],
-	outputs=[keyword],
+	outputs=[labels],
 	examples=[
 		['backward16k.wav'],
 		['happy16k.wav'],
@@ -35,6 +31,7 @@ iface = gr.Interface(
 		['stop16k.wav'],
 		['up16k.wav'],
 	],
-    allow_flagging=False
+    allow_flagging='never',
 )
+
 iface.launch()