|
import gradio as gr |
|
import torch |
|
from datasets import load_dataset |
|
from transformers import pipeline, SpeechT5Processor, SpeechT5HifiGan, SpeechT5ForTextToSpeech |
|
|
|
model_id = "Sandiago21/speecht5_finetuned_voxpopuli_it" |
|
|
|
model = SpeechT5ForTextToSpeech.from_pretrained(model_id) |
|
vocoder = SpeechT5HifiGan.from_pretrained("microsoft/speecht5_hifigan") |
|
embeddings_dataset = load_dataset("Matthijs/cmu-arctic-xvectors", split="validation") |
|
speaker_embeddings = torch.tensor(embeddings_dataset[7440]["xvector"]).unsqueeze(0) |
|
|
|
checkpoint = "microsoft/speecht5_tts" |
|
processor = SpeechT5Processor.from_pretrained(checkpoint) |
|
|
|
def synthesize_speech(text): |
|
inputs = processor(text=text, return_tensors="pt") |
|
|
|
speech = model.generate_speech(inputs["input_ids"], speaker_embeddings, vocoder=vocoder) |
|
|
|
return gr.Audio.update(value=(16000, speech.cpu().numpy())) |
|
|
|
syntesize_speech_gradio = gr.Interface( |
|
synthesize_speech, |
|
inputs = gr.Textbox(label="Text", placeholder="Type something here..."), |
|
outputs=gr.Audio(), |
|
|
|
).launch() |
|
|
|
|