| import torch | |
| from transformers import SpeechT5Processor, SpeechT5ForTextToSpeech | |
| from datasets import load_dataset | |
| import soundfile as sf | |
| MODEL_ID = "microsoft/speecht5_tts" | |
| processor = SpeechT5Processor.from_pretrained(MODEL_ID) | |
| model = SpeechT5ForTextToSpeech.from_pretrained(MODEL_ID) | |
| vocoder = torch.hub.load("snakers4/silero-vad", "silero_vad", force_reload=True) | |
| def synthesize_speech(text_input): | |
| inputs = processor(text=text_input, return_tensors="pt") | |
| with torch.no_grad(): | |
| speech = model.generate_speech(inputs["input_ids"], vocoder=vocoder) | |
| sf.write("output.wav", speech.numpy(), 16000) | |
| return "output.wav" | |