Update app.py

app.py

@@ -2,26 +2,33 @@ import gradio as gr
 from faster_whisper import WhisperModel
 
 # Load the Faster Whisper model
-model_name = "Systran/faster-whisper-large-v3"
-model = WhisperModel(model_name, device="cpu")  # Use "cuda" if you have a GPU
+model = WhisperModel("large-v3", device="cpu")  # Use "cuda" for GPU
 
-# Define a transcription function
+# Define the transcription function
 def transcribe_audio(audio_file):
+    """
+    Transcribes the audio file using the Faster Whisper model.
+    """
     try:
-        segments, info = model.transcribe(audio_file, beam_size=5)  # Customize parameters as needed
-        transcription = " ".join([segment.text for segment in segments])
+        segments, info = model.transcribe(audio_file, beam_size=5)  # Adjust beam_size as needed
+        transcription = "\n".join(
+            [f"[{segment.start:.2f}s -> {segment.end:.2f}s] {segment.text}" for segment in segments]
+        )
         return transcription
     except Exception as e:
         return f"Error: {str(e)}"
 
-# Create Gradio interface
+# Create the Gradio interface
 interface = gr.Interface(
-    fn=transcribe_audio,
-    inputs=gr.Audio(source="upload", type="filepath", label="Upload Audio"),
-    outputs=gr.Textbox(label="Transcription"),
-    title="Sinhala Audio-to-Text Transcription",
-    description="Upload an audio file and get the transcription in Sinhala using the Faster Whisper model.",
-    allow_flagging="never"
+    fn=transcribe_audio,  # Function to process the input
+    inputs=gr.Audio(source="upload", type="filepath", label="Upload Audio"),  # Input: Audio file
+    outputs=gr.Textbox(label="Transcription"),  # Output: Textbox for the transcription
+    title="Audio-to-Text Transcription",
+    description=(
+        "Upload an audio file and get the transcription using the Faster Whisper model "
+        "large-v3. Supports high-quality transcription with beam search."
+    ),
+    allow_flagging="never",
 )
 
 # Launch the Gradio app