code simplified

app.py

@@ -1,15 +1,11 @@
 import os
-import math
-import gradio as gr
 import torch
 import librosa
-import pandas as pd
 import numpy as np
-
+import gradio as gr
 from sonics import HFAudioClassifier
 
-
-# Constants
+# Model configurations
 MODEL_IDS = {
     "SpecTTTra-α (5s)": "awsaf49/sonics-spectttra-alpha-5s",
     "SpecTTTra-β (5s)": "awsaf49/sonics-spectttra-beta-5s",
@@ -19,11 +15,9 @@ MODEL_IDS = {
     "SpecTTTra-γ (120s)": "awsaf49/sonics-spectttra-gamma-120s",
 }
 
-
 device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
 model_cache = {}
 
-
 def load_model(model_name):
     """Load model if not already cached"""
     if model_name not in model_cache:
@@ -34,21 +28,14 @@ def load_model(model_name):
         model_cache[model_name] = model
     return model_cache[model_name]
 
-
 def process_audio(audio_path, model_name):
     """Process audio file and return prediction"""
     try:
-        # Load model
         model = load_model(model_name)
-
-        # Get max time from model config
         max_time = model.config.audio.max_time
-
+        
         # Load and process audio
         audio, sr = librosa.load(audio_path, sr=16000)
-        duration = len(audio) / sr
-
-        # Calculate chunk size and middle position
         chunk_samples = int(max_time * sr)
         total_chunks = len(audio) // chunk_samples
         middle_chunk_idx = total_chunks // 2
@@ -57,89 +44,74 @@ def process_audio(audio_path, model_name):
         start = middle_chunk_idx * chunk_samples
         end = start + chunk_samples
         chunk = audio[start:end]
-
-        # Pad if needed (shouldn't be necessary for middle chunk)
+        
         if len(chunk) < chunk_samples:
             chunk = np.pad(chunk, (0, chunk_samples - len(chunk)))
-
-        # Convert to tensor and get prediction
+            
+        # Get prediction
         with torch.no_grad():
             chunk = torch.from_numpy(chunk).float().to(device)
             pred = model(chunk.unsqueeze(0))
             prob = torch.sigmoid(pred).cpu().numpy()[0]
-
-        # Get prediction
-        output = {"Real": 1 - prob, "Fake": prob}
-
-        return output
+            
+        return {"Real": 1 - prob, "Fake": prob}
 
     except Exception as e:
-        return {
-            "Duration": "Error",
-            "Prediction": f"Error: {str(e)}",
-            "Confidence": "N/A",
-        }
-
+        return {"Error": str(e)}
 
 def predict(audio_file, model_name):
     """Gradio interface function"""
     if audio_file is None:
-        return {
-            "Duration": "No file",
-            "Prediction": "Please upload an audio file",
-            "Confidence": "N/A",
-        }
-
+        return {"Message": "Please upload an audio file"}
     return process_audio(audio_file, model_name)
 
-
 # Create Gradio interface
-css = """
-.heading {
-    text-align: center;
-    margin-bottom: 2rem;
-}
-.logo {
-    max-width: 250px;
-    margin: 0 auto;
-    display: block;
-}
-"""
-
-with gr.Blocks(css=css) as demo:
+with gr.Blocks() as demo:
     gr.HTML(
         """
-        <div class="heading">
-            <img src="https://i.postimg.cc/3Jx3yZ5b/real-vs-fake-sonics-w-logo.jpg" class="logo">
+        <div style="text-align: center; margin-bottom: 1rem;">
+            <img src="https://i.postimg.cc/3Jx3yZ5b/real-vs-fake-sonics-w-logo.jpg" 
+                 style="max-width: 300px; margin: 0 auto;">
             <h1>SONICS: Synthetic Or Not - Identifying Counterfeit Songs</h1>
-            <h3><span style="color:red;"><b>ICLR 2025 [Poster]</b></span></h3>
+            <h3>ICLR 2025 [Poster]</h3>
         </div>
-    """
+        """
     )
-
+    
     with gr.Row():
         with gr.Column():
-            audio_input = gr.Audio(label="Upload Audio", type="filepath")
+            audio_input = gr.Audio(
+                label="Upload Audio File",
+                type="filepath"
+            )
             model_dropdown = gr.Dropdown(
                 choices=list(MODEL_IDS.keys()),
                 value="SpecTTTra-γ (5s)",
-                label="Select Model",
+                label="Select Model"
             )
-            submit_btn = gr.Button("Predict")
-
+            submit_btn = gr.Button("Analyze Audio")
+        
         with gr.Column():
-            output = gr.Label(label="Result", num_top_classes=2)
-
-    submit_btn.click(fn=predict, inputs=[audio_input, model_dropdown], outputs=[output])
-
+            output = gr.Label(
+                label="Analysis Result",
+                num_top_classes=2
+            )
+    
     gr.Markdown(
         """
-    ## Resources
-    - 📄 [Paper](https://openreview.net/forum?id=PY7KSh29Z8)
-    - 🎵 [Dataset](https://huggingface.co/datasets/awsaf49/sonics)
-    - 🔬 [ArXiv](https://arxiv.org/abs/2408.14080)
-    - 💻 [GitHub](https://github.com/awsaf49/sonics)
-    """
+        ### Resources
+        - [📄 Paper](https://openreview.net/forum?id=PY7KSh29Z8)
+        - [🎵 Dataset](https://huggingface.co/datasets/awsaf49/sonics)
+        - [🔬 ArXiv](https://arxiv.org/abs/2408.14080)
+        - [💻 GitHub](https://github.com/awsaf49/sonics)
+        """
+    )
+    
+    submit_btn.click(
+        fn=predict,
+        inputs=[audio_input, model_dropdown],
+        outputs=[output]
     )
 
-demo.launch()
+if __name__ == "__main__":
+    demo.launch()