Update app.py

app.py

@@ -73,14 +73,10 @@ print("----------> Loaded models <-----------")
 
 gpu_timeout = int(os.getenv("GPU_TIMEOUT", 60))
 @spaces.GPU(duration=gpu_timeout)
-def generator(microphone, file_upload, num_speakers, max_duration, history):
+def generator(file_upload, num_speakers, max_duration, history):
     history = history or ""
     
-    if microphone:
-        path = microphone
-    elif file_upload:
-        path = file_upload
-    
+    path = file_upload
     waveform, sampling_rate = librosa.load(path, sr=SAMPLE_RATE, mono=True, duration=max_duration)
 
     print(waveform.shape, sampling_rate)
@@ -132,23 +128,45 @@ def generator(microphone, file_upload, num_speakers, max_duration, history):
     yield history, history, file_name
 
 
+# demo = gr.Interface(
+#     generator, 
+#     inputs=[
+#         gr.Audio(type="filepath"),
+#         gr.Number(value=1, label="Number of Speakers"),
+#         gr.Number(value=120, label="Maximum Duration (Seconds)"),
+#         'state',
+#     ],
+#     outputs=['text', 'state', 'file'],
+#     title="Mongolian Whisper 🇲🇳",
+#     description=(
+#         "Transcribe Microphone / Uploaded File in Mongolian Whisper Model."
+#     )
+# )
+
+import gradio as gr
+
 demo = gr.Interface(
-    generator, 
-    inputs=[
-        gr.Audio(type="filepath"),
-        gr.Audio(type="filepath"),
-        gr.Number(value=1, label="Number of Speakers"),
-        gr.Number(value=120, label="Maximum Duration (Seconds)"),
-        'state',
-    ],
-    outputs=['text', 'state', 'file'],
-    title="Mongolian Whisper 🇲🇳",
-    description=(
-        "Transcribe Microphone / Uploaded File in Mongolian Whisper Model."
+    fn=generator, 
+    inputs=gr.Column(  # 👈 wrap inputs in a Column
+        [
+            gr.Audio(type="filepath"),
+            gr.Number(value=1, label="Number of Speakers within Audio."),
+            gr.Number(value=120, label="Maximum Duration (Seconds)"),
+            gr.State(),  # 'state' should be a gr.State component, not a plain string
+        ]
     ),
-    allow_flagging="never",
+    outputs=gr.Column(  # 👈 wrap outputs in a Column
+        [
+            gr.Textbox(label="Transcription"),
+            gr.State(),
+            gr.File(label="Output File"),
+        ]
+    ),
+    title="Mongolian Whisper 🇲🇳",
+    description="Transcribe Microphone / Uploaded File in Mongolian Whisper Model."
 )
 
+
 # define queue - required for generators
 demo.queue()