try adding librosa

app.py

@@ -1,8 +1,15 @@
 import gradio as gr
+import librosa
+
+def load_audio_to_tensor(filename):
+    audio, sampling_rate = librosa.load(filename, sr=None, mono=True)  # load audio and convert to mono
+    wave = librosa.resample(audio, orig_sr=sampling_rate, target_sr=16000)  # resample to 16KHz
+    return wave
 
 def greet(name):
+    load_audio_to_tensor(name)
     # return "Hello " + str(name[1]) + "!!"
-    return name
+    return name + 'type'
 
 iface = gr.Interface(fn=greet, inputs="file", outputs="text")
 # iface = gr.Interface(fn=greet, inputs="audio", outputs="text")

requirements.txt

@@ -0,0 +1 @@
+librosa==0.10.1