Update app

app.py

@@ -1,14 +1,13 @@
-import gradio as gr
-import time
-import urllib.request
-from pathlib import Path
 import os
+import time
 import torch
-import scipy.io.wavfile
+import urllib.request
+import gradio as gr
 import numpy as np
+import soundfile as sf
 from espnet2.bin.tts_inference import Text2Speech
 from espnet2.utils.types import str_or_none
-
+from pathlib import Path
 
 gos_text2speech = Text2Speech.from_pretrained(
     model_tag="bartelds/gos_tts",
@@ -22,13 +21,13 @@ def inference(text,lang):
   with torch.no_grad():
       if lang == "Hoogelaandsters":
           wav = gos_text2speech(text.lower(), sids=np.array([1]))["wav"]
-          scipy.io.wavfile.write("out.wav", gos_text2speech.fs , wav.view(-1).cpu().numpy())
+          sf.write("out.wav", wav.view(-1).cpu().numpy(), gos_text2speech.fs)
       if lang == "Oldambsters":
           wav = gos_text2speech(text.lower(), sids=np.array([2]))["wav"]
-          scipy.io.wavfile.write("out.wav", gos_text2speech.fs , wav.view(-1).cpu().numpy())
+          sf.write("out.wav", wav.view(-1).cpu().numpy(), gos_text2speech.fs)
       if lang == "Westerkertaaiers":
          wav = gos_text2speech(text.lower(), sids=np.array([3]))["wav"]
-         scipy.io.wavfile.write("out.wav", gos_text2speech.fs , wav.view(-1).cpu().numpy())
+         sf.write("out.wav", wav.view(-1).cpu().numpy(), gos_text2speech.fs)
 
   return  "out.wav", "out.wav"