Spaces:
Runtime error
Runtime error
File size: 3,474 Bytes
b99241b b230689 b99241b b230689 b99241b b230689 b99241b b230689 b99241b b230689 b99241b b230689 b99241b b230689 b99241b b230689 b99241b b230689 b99241b b230689 b99241b 08a9a34 b99241b 240f8ad 3235a32 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 |
import gradio as gr
import numpy as np
import torch
from datasets import load_dataset
from transformers import (
SpeechT5ForTextToSpeech,
SpeechT5HifiGan,
SpeechT5Processor,
pipeline,
VitsModel,
VitsTokenizer,
)
device = "cuda:0" if torch.cuda.is_available() else "cpu"
# load speech translation checkpoint
asr_pipe = pipeline(
"automatic-speech-recognition", model="openai/whisper-base", device=device
)
# speecht5
# load text-to-speech checkpoint and speaker embeddings
# 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)
# embeddings_dataset = load_dataset("Matthijs/cmu-arctic-xvectors", split="validation")
# speaker_embeddings = torch.tensor(embeddings_dataset[7306]["xvector"]).unsqueeze(0)
# mms
model = VitsModel.from_pretrained("Matthijs/mms-tts-fra")
tokenizer = VitsTokenizer.from_pretrained("Matthijs/mms-tts-fra")
# 保持 main 函数 speech_to_speech_translation 不变
# 并根据需要仅更新 translate 和 synthesise 函数
def translate(audio):
# outputs = asr_pipe(audio, max_new_tokens=256, generate_kwargs={"task": "translate"})
outputs = asr_pipe(
audio,
max_new_tokens=256,
generate_kwargs={"task": "transcribe", "language": "fr"},
# generate_kwargs={"task": "transcribe"},
)
print(outputs)
return outputs["text"]
# speecht5
# 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 = tokenizer(text, return_tensors="pt")
input_ids = inputs["input_ids"]
with torch.no_grad():
outputs = model(input_ids)
speech = outputs.audio[0]
return speech.cpu()
def speech_to_speech_translation(audio):
translated_text = translate(audio)
synthesised_speech = synthesise(translated_text)
synthesised_speech = (synthesised_speech.numpy() * 32767).astype(np.int16)
return 16000, synthesised_speech
title = "Cascaded STST"
description = """
Demo for cascaded speech-to-speech translation (STST), mapping from source speech in any language to target speech in Chinese. Demo uses OpenAI's [Whisper Base](https://huggingface.co/openai/whisper-base) model for speech translation, and Microsoft's
[MMS TTS](https://huggingface.co/Matthijs/mms-tts-fra) model for text-to-speech:
"""
demo = gr.Blocks()
mic_translate = gr.Interface(
fn=speech_to_speech_translation,
inputs=gr.Audio(source="microphone", type="filepath"),
outputs=gr.Audio(label="Generated Speech", type="numpy"),
title=title,
description=description,
)
file_translate = gr.Interface(
fn=speech_to_speech_translation,
inputs=gr.Audio(source="upload", type="filepath"),
outputs=gr.Audio(label="Generated Speech", type="numpy"),
examples=[["./example.wav"]],
title=title,
description=description,
)
with demo:
gr.TabbedInterface([mic_translate, file_translate], ["Microphone", "Audio File"])
# demo.launch(share=True)
demo.launch()
|