Update app.py

app.py

@@ -8,7 +8,7 @@ from datasets import Dataset, DatasetDict
 from datasets import load_dataset
 from df.enhance import enhance, init_df, load_audio, save_audio
 model_enhance, df_state, _ = init_df()
-def remove_nn(wav, sample_rate=44100):
+def remove_nn(wav, sample_rate=16000):
     
     audio=librosa.resample(wav,orig_sr=sample_rate,target_sr=df_state.sr(),)
     
@@ -21,7 +21,7 @@ def remove_nn(wav, sample_rate=44100):
    # save_audio("enhanced.wav", enhanced, df_state.sr())
     audiodata=librosa.resample(enhanced[0].numpy(),orig_sr=df_state.sr(),target_sr=sample_rate)
    
-    return 44100, audiodata/np.max(audiodata)
+    return 16000, audiodata/np.max(audiodata)
 class DataEditor:
     def __init__(self,df):
         self.df=df
@@ -120,7 +120,7 @@ class DataEditor:
             txt_audio = row['text']
             row_audio = self.sdata[self.current_selected]
             if row['flag'] !=0:
-                return txt_audio,(44100,row_audio)
+                return txt_audio,(16000,row_audio)
             else :
                 return txt_audio,None
           
@@ -137,8 +137,13 @@ class DataEditor:
                 #row['text'] = text
                 row['flag']=1
                 self.data.iloc[self.current_selected] = row
-                sr,audio_array=data_oudio
-                self.sdata[self.current_selected] =audio_array
+                sr,audio=data_oudio
+                if sr!=16000:
+                    audio=audio.astype(np.float32)
+                    audio/=np.max(np.abs(audio))
+                    audio=librosa.resample(audio,orig_sr=sr,target_sr=16000)
+                self.sdata[self.current_selected] =audio
+               
         return  "",None
     def on_row_delete(self):
         if self.current_selected>=0: