Update description

app.py

@@ -5,6 +5,7 @@ import numpy as np
 import torchaudio
 import librosa
 
+
 import gradio as gr
 from modules import load_audio, MosPredictor, denorm
 
@@ -26,7 +27,13 @@ model_asli = model_asli.to(device)
 
 
 def predict_mos(wavefile:str):
-
+    device = "cuda:0" if torch.cuda.is_available() else "cpu"
+    if device != model.device:
+        model.to(device)
+    if device != model_asli.device:
+        model_asli.to(device)
+    
+    
     print('Starting prediction...')
     # STFT
     wav = torchaudio.load(wavefile)[0] 
@@ -74,8 +81,10 @@ title =  """
 """ 
 
 description = """
-This is a demo of [MOSA-Net+](https://github.com/dhimasryan/MOSA-Net-Cross-Domain/tree/main/MOSA_Net%2B), 
-an enhanced version of the multi-objective speech assessment model MOSA-Net, by leveraging the acoustic features from Whisper, a large-scaled weakly supervised model. 
+This is a demo of [MOSA-Net+](https://github.com/dhimasryan/MOSA-Net-Cross-Domain/tree/main/MOSA_Net%2B), an improved version of MOSA-
+NET that predicts human-based speech quality and intelligibility. MOSA-Net+ uses Whisper to generate cross-domain features. The model employs a CNN-
+BLSTM architecture with an attention mechanism and is trained using a multi-task learning approach to predict subjective listening test
+scores.  
 MOSA-Net+ was tested in the noisy-and-enhanced track of the VoiceMOS Challenge 2023, where it obtained the top-ranked performance among nine systems [full paper](https://arxiv.org/abs/2309.12766)
 """