Update app.py

app.py

@@ -5,14 +5,15 @@ import torch
 import numpy as np
 import torchaudio
 
+model = VitsModel.from_pretrained("SeyedAli/Persian-Speech-synthesis")
+tokenizer = AutoTokenizer.from_pretrained("SeyedAli/Persian-Speech-synthesis")
 
 def TTS(text):
-   model = VitsModel.from_pretrained("SeyedAli/Persian-Speech-synthesis")
-   tokenizer = AutoTokenizer.from_pretrained("SeyedAli/Persian-Speech-synthesis")
    inputs = tokenizer(text, return_tensors="pt")
-   pipe = pipeline("text-to-speech", model=model,tokenizer=tokenizer)
+   with torch.no_grad():
+        output = model(**inputs).waveform 
    with tempfile.NamedTemporaryFile(suffix=".wav", delete=False) as fp:
-        torchaudio.save(fp, pipe(text)['audio'], pipe(text)['sampling_rate'])
+        torchaudio.save(fp, output, model.config.sampling_rate)
         return fp.name
 iface = gr.Interface(fn=TTS, inputs="text", outputs="audio")
 iface.launch(share=False)