Update app.py

app.py

@@ -5,9 +5,9 @@ import os
 model_id = "HarshitJoshi/whisper-small-Hindi"
 pipe = pipeline("automatic-speech-recognition", model=model_id)
 
-def transcribe_speech(filepath):
+def transcribe_speech(audio):
     output = pipe(
-        filepath,
+        audio,
         max_new_tokens=256,
         generate_kwargs={
             "task": "transcribe",
@@ -21,25 +21,40 @@ def transcribe_speech(filepath):
 example_folder = "./examples"
 example_files = [f for f in os.listdir(example_folder) if f.endswith('.wav') or f.endswith('.mp3')]
 
-def handle_input(mic, upload, example):
-    if mic is not None:
-        return transcribe_speech(mic)
-    elif upload is not None:
-        return transcribe_speech(upload.name)
-    elif example is not None:
-        return transcribe_speech(os.path.join(example_folder, example))
-    else:
-        return "Please provide an input."
+def play_and_transcribe(filename):
+    filepath = os.path.join(example_folder, filename)
+    transcription = transcribe_speech(filepath)
+    return filepath, transcription
 
 with gr.Blocks() as demo:
-    with gr.Row():
-        mic = gr.Audio(type="filepath", label="Record from Microphone")
-        upload = gr.Audio(type="file", label="Upload an Audio File")
-        example = gr.Dropdown(choices=example_files, label="Or Select an Example")
-
-    output = gr.Textbox(label="Transcription")
+    gr.Markdown("# Hindi Speech Transcription")
     
-    submit_btn = gr.Button("Transcribe")
-    submit_btn.click(handle_input, inputs=[mic, upload, example], outputs=output)
+    with gr.Tab("Transcribe"):
+        audio_input = gr.Audio(source="microphone", type="filepath", label="Audio Input")
+        upload_button = gr.UploadButton("Upload Audio", file_types=["audio"])
+        transcribe_button = gr.Button("Transcribe")
+        output_text = gr.Textbox(label="Transcription")
+
+        transcribe_button.click(
+            fn=transcribe_speech,
+            inputs=audio_input,
+            outputs=output_text
+        )
+        upload_button.upload(
+            fn=lambda file: file.name,
+            inputs=upload_button,
+            outputs=audio_input
+        )
+
+    with gr.Tab("Examples"):
+        example_dropdown = gr.Dropdown(choices=example_files, label="Select an example")
+        example_audio = gr.Audio(label="Audio Playback")
+        example_transcription = gr.Textbox(label="Transcription")
+
+        example_dropdown.change(
+            fn=play_and_transcribe,
+            inputs=example_dropdown,
+            outputs=[example_audio, example_transcription]
+        )
 
-demo.launch(debug=True)
+demo.launch(debug=True)