Update app.py

app.py

@@ -1,40 +1,34 @@
+# Import necessary libraries
 import gradio as gr
-import torch
 import torchaudio
 from audioseal import AudioSeal
-from io import BytesIO
-
-def watermark_audio(file_info):
-    audio, sr = torchaudio.load(file_info)
-    model = AudioSeal.load_generator("audioseal_wm_16bits")
-    audios = audio.unsqueeze(0)  # Add batch dimension
-    watermarked_audio = model(audios, alpha=1)
-    output = BytesIO()
-    torchaudio.save(output, watermarked_audio.squeeze(0), sr)
-    output.seek(0)
-    return output, 'audio/wav'
+import torch
 
+# Function to detect watermarks in the audio and determine if it's AI-generated
 def detect_watermark(file_info):
+    # Load the audio file from the uploaded file
     audio, sr = torchaudio.load(file_info)
-    audios = audio.unsqueeze(0)  # Add batch dimension
+    audio = audio.unsqueeze(0)  # Add batch dimension as expected by the AudioSeal detector
+
+    # Initialize the AudioSeal detector
     detector = AudioSeal.load_detector("audioseal_detector_16bits")
-    result, message = detector.detect_watermark(audios, message_threshold=0.5)
-    return f"Watermark Detected: {result}, Message: {message.numpy()}"
 
-with gr.Blocks() as demo:
-    with gr.Row():
-        with gr.Column():
-            gr.Markdown("### Upload Audio for Watermarking")
-            audio_input = gr.Audio(label="Upload Audio", type="file")
-            watermark_btn = gr.Button("Watermark Audio")
-            watermarked_audio_output = gr.Audio(label="Download Watermarked Audio", type="numpy", autoplay=True)
-            watermark_btn.click(watermark_audio, inputs=audio_input, outputs=watermarked_audio_output)
-        with gr.Column():
-            gr.Markdown("### Upload Audio for Watermark Detection")
-            audio_input_detect = gr.Audio(label="Upload Audio", type="file")
-            detect_btn = gr.Button("Detect Watermark")
-            detection_output = gr.Textbox(label="Detection Result")
-            detect_btn.click(detect_watermark, inputs=audio_input_detect, outputs=detection_output)
+    # Detect watermarks in the audio
+    result, message = detector.detect_watermark(audio, message_threshold=0.5)
+    
+    # Interpret the detection result
+    detection_result = "AI-generated" if result else "genuine"
+    return f"This audio is likely {detection_result} based on watermark detection."
+
+# Define Gradio interface
+interface = gr.Interface(fn=detect_watermark,
+                         inputs=gr.Audio(source="upload", type="file", label="Upload your audio"),
+                         outputs="text",
+                         title="Deep Fake Defender: AI Voice Cloning Detection",
+                         description="Upload an audio file to check if it's AI-generated or genuine.")
+
+# Launch the Gradio app
+if __name__ == "__main__":
+    interface.launch()
 
-demo.launch()