Update app.py

app.py

@@ -60,10 +60,7 @@ def generate(image, video, audio, instruction=DEFAULT_INPUT, sampling=False, tem
         vid = encode_video(video)
         inputs = [{"role": "user", "content": vid + [instruction]}]
     elif audio is not None:
-        if isinstance(audio, str):
-            aud, _ = librosa.load(audio, sr=16000, mono=True)
-        else:
-            aud = audio
+        aud, _ = librosa.load(audio, sr=16000, mono=True)
         inputs = [{"role": "user", "content": [aud, instruction]}]
     else:
         return "No input provided."
@@ -94,7 +91,7 @@ with gr.Blocks(css=css) as main:
     with gr.Column():
         input = gr.Image(label="Image")
         input_2 = gr.Video(label="Video")
-        input_3 = gr.Audio(label="Audio")
+        input_3 = gr.Audio(label="Audio", type="filepath")
         instruction = gr.Textbox(lines=1, value=DEFAULT_INPUT, label="Instruction")
         sampling = gr.Checkbox(value=False, label="Sampling")
         temperature = gr.Slider(minimum=0.01, maximum=1.99, step=0.01, value=0.7, label="Temperature")