microsoft

/

speecht5_tts

like 520

from transformers import SpeechT5Processor, SpeechT5ForTextToSpeech, SpeechT5HifiGan
from datasets import load_dataset
import torch
import soundfile as sf
from datasets import load_dataset

processor = SpeechT5Processor.from_pretrained("microsoft/speecht5_tts")
model = SpeechT5ForTextToSpeech.from_pretrained("microsoft/speecht5_tts")
vocoder = SpeechT5HifiGan.from_pretrained("microsoft/speecht5_hifigan")

inputs = processor(text="Hello, my dog is cute.", return_tensors="pt")

# embeddings_dataset = load_dataset("Matthijs/cmu-arctic-xvectors", split="validation")
# speaker_embeddings = torch.tensor(embeddings_dataset[7306]["xvector"]).unsqueeze(0)

speech = model.generate_speech(inputs["input_ids"], vocoder=vocoder)#, speaker_embeddings,)

sf.write("speech.wav", speech.numpy(), samplerate=16000)

I think gonna give the output from a llm to this model when the llm yeild the text in real time how can we make this read in real time?