Update app.py

app.py

@@ -4,7 +4,6 @@ import torch
 from transformers import pipeline, VitsModel, AutoTokenizer, AutoTokenizer
 from transformers import SpeechT5ForTextToSpeech, SpeechT5HifiGan, SpeechT5Processor
 from transformers import WhisperTokenizer, GenerationConfig
-#from transformers import BarkModel, AutoProcessor
 
 
 device = "cuda:0" if torch.cuda.is_available() else "cpu"
@@ -39,17 +38,6 @@ speaker_embeddings2 = np.load('speaker_embeddings.npy')
 speaker_embeddings2 = torch.tensor(speaker_embeddings2)
 print(speaker_embeddings2)
 
-#lang_detector = pipeline("text-classification", model="papluca/xlm-roberta-base-language-detection")
-
-# ---------------- Speech generator  bark--------------------------#
-
-
-#model = BarkModel.from_pretrained("suno/bark-small")
-#processor = BarkProcessor.from_pretrained("suno/bark-small")
-
-# processor = AutoProcessor.from_pretrained("suno/bark-small")
-# model = BarkModel.from_pretrained("suno/bark-small")
-
 
 def language_detector(text):
     resultado = lang_detector(text)
@@ -62,23 +50,6 @@ def translate(audio):
     print(outputs["text"])
     return outputs["text"]
 
-
-# def synthesise(text):
-#     inputs = processor(text=text, voice_preset="v2/es_speaker_8")
-#     speech_output = model.generate(**inputs).cpu()
-#     return speech_output
-
-# def speech_to_speech_translation(audio):
-#     translated_text = translate(audio)
-#     synthesised_speech = synthesise(translated_text)
-    
-#     sample_rate = model.generation_config.sample_rate
-
-#     synthesised_speech = synthesised_speech.numpy().squeeze()
-    
-#     return sample_rate, synthesised_speech
-
-    
 def synthesise(text):
     inputs = processor(text=text, return_tensors="pt")
     output = model.generate_speech(inputs["input_ids"], speaker_embeddings2, vocoder=vocoder)