Create app.py

app.py

@@ -0,0 +1,36 @@
+import gradio as gr
+from transformers import pipeline
+import torch
+
+# Initialize the transcriber
+def initialize_transcriber():
+    return pipeline("automatic-speech-recognition", 
+                   model="vinai/PhoWhisper-medium",
+                   device="cuda" if torch.cuda.is_available() else "cpu")
+
+transcriber = initialize_transcriber()
+
+# Function to transcribe audio
+def transcribe_audio(audio_path):
+    try:
+        # Transcribe the audio
+        result = transcriber(audio_path)
+        transcribed_text = result["text"]
+        return transcribed_text
+    except Exception as e:
+        return f"Error during transcription: {str(e)}"
+
+# Create the Gradio interface
+interface = gr.Interface(
+    fn=transcribe_audio,
+    inputs=gr.Audio(source="microphone", type="filepath"),
+    outputs="text",
+    title="Vietnamese Speech-to-Text",
+    description="Record audio in Vietnamese and get the transcription",
+    examples=[],
+    theme=gr.themes.Soft()
+)
+
+# Launch the app
+if __name__ == "__main__":
+    interface.launch(share=True)