Update app.py

app.py

@@ -21,10 +21,10 @@ en_fr_translator = pipeline("translation_en_to_fr")
 
 
 # load text-to-speech
-model = VitsModel.from_pretrained("facebook/mms-tts-fra")
+model_new = VitsModel.from_pretrained("facebook/mms-tts-fra")
 tokenizer = VitsTokenizer.from_pretrained("facebook/mms-tts-fra")
 
-def synthesise(text):
+def synthesise_new(text):
     translation_to_french = en_fr_translator(text)
     french_text = translation_to_french[0]['translation_text']
 
@@ -32,25 +32,25 @@ def synthesise(text):
     input_ids = inputs["input_ids"]
     
     with torch.no_grad():
-        outputs = model(input_ids)
+        outputs = model_new(input_ids)
     
     speech = outputs["waveform"]
     return speech
 
 
 # load text-to-speech checkpoint and speaker embeddings
-#processor = SpeechT5Processor.from_pretrained("microsoft/speecht5_tts")
+processor = SpeechT5Processor.from_pretrained("microsoft/speecht5_tts")
 
-#model = SpeechT5ForTextToSpeech.from_pretrained("microsoft/speecht5_tts").to(device)
-#vocoder = SpeechT5HifiGan.from_pretrained("microsoft/speecht5_hifigan").to(device)
+model = SpeechT5ForTextToSpeech.from_pretrained("microsoft/speecht5_tts").to(device)
+vocoder = SpeechT5HifiGan.from_pretrained("microsoft/speecht5_hifigan").to(device)
 
-#embeddings_dataset = load_dataset("Matthijs/cmu-arctic-xvectors", split="validation")
-#speaker_embeddings = torch.tensor(embeddings_dataset[7306]["xvector"]).unsqueeze(0)
+embeddings_dataset = load_dataset("Matthijs/cmu-arctic-xvectors", split="validation")
+speaker_embeddings = torch.tensor(embeddings_dataset[7306]["xvector"]).unsqueeze(0)
 
-#def synthesise(text):
-#    inputs = processor(text=text, return_tensors="pt")
-#    speech = model.generate_speech(inputs["input_ids"].to(device), speaker_embeddings.to(device), vocoder=vocoder)
-#    return speech.cpu()
+def synthesise(text):
+    inputs = processor(text=text, return_tensors="pt")
+    speech = model.generate_speech(inputs["input_ids"].to(device), speaker_embeddings.to(device), vocoder=vocoder)
+    return speech.cpu()
 
 
 def speech_to_speech_translation(audio):