Spaces:
Build error
Build error
File size: 2,963 Bytes
cf18f93 7f71b9c d8fc388 04c250d f6fdf7a 04c250d 58c3446 04c250d f6fdf7a f576d11 3cefd18 04c250d 970053f 576bebe f4b46dd 04c250d 293d0d2 f6fdf7a 970053f 576bebe 9be2afa c2a1b9b 32fedd3 f4b46dd f6fdf7a d3fe331 04c250d f6fdf7a 04c250d ee620af 04c250d f6fdf7a 970053f f6fdf7a 04c250d f6fdf7a |
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 |
import os
os.environ["CURL_CA_BUNDLE"]=""
import gradio as gr
import time
import urllib.request
from pathlib import Path
import os
import torch
import scipy.io.wavfile
from espnet2.bin.tts_inference import Text2Speech
from espnet2.utils.types import str_or_none
from parallel_wavegan.utils import download_pretrained_model
gos_text2speech = Text2Speech.from_pretrained(
model_tag="https://huggingface.co/ahnafsamin/FastSpeech2-gronings/resolve/main/tts_train_fastspeech2_raw_char_tacotron_train.loss.ave.zip",
vocoder_tag="parallel_wavegan/ljspeech_parallel_wavegan.v3"
)
hoogelandsters_text2speech = Text2Speech.from_pretrained(
model_tag="https://huggingface.co/ahnafsamin/FastSpeech2-gronings-hoogelandsters/resolve/main/tts_train_fastspeech2_raw_char_tacotron_train.loss.ave.zip",
vocoder_tag="parallel_wavegan/ljspeech_parallel_wavegan.v3"
)
westerkwartiers_text2speech = Text2Speech.from_pretrained(
model_tag="https://huggingface.co/ahnafsamin/FastSpeech2-gronings-westerkwartiers/resolve/main/tts_train_fastspeech2_raw_char_tacotron_train.loss.ave.zip",
vocoder_tag="parallel_wavegan/ljspeech_parallel_wavegan.v3"
)
oldambster_text2speech = Text2Speech.from_pretrained(
model_tag="https://huggingface.co/ahnafsamin/FastSpeech2-gronings-oldambster/resolve/main/tts_train_fastspeech2_raw_char_tacotron_train.loss.ave.zip",
vocoder_tag="parallel_wavegan/ljspeech_parallel_wavegan.v3"
)
def inference(text,lang):
with torch.no_grad():
if lang == "gronings":
wav = gos_text2speech(text)["wav"]
scipy.io.wavfile.write("out.wav", gos_text2speech.fs , wav.view(-1).cpu().numpy())
elif lang == "gronings hoogelandsters":
wav = hoogelandsters_text2speech(text)["wav"]
scipy.io.wavfile.write("out.wav", hoogelandsters_text2speech.fs , wav.view(-1).cpu().numpy())
elif lang == "gronings westerkwartiers":
wav = westerkwartiers_text2speech(text)["wav"]
wav = wav * 15
scipy.io.wavfile.write("out.wav", westerkwartiers_text2speech.fs , wav.view(-1).cpu().numpy())
#data, sr = librosa.load("output.wav")
#factor = 2.0
#data *= factor
#sf.write("out.wav", data, sr)
elif lang == "gronings oldambster":
wav = oldambster_text2speech(text)["wav"]
scipy.io.wavfile.write("out.wav", oldambster_text2speech.fs , wav.view(-1).cpu().numpy())
return "out.wav", "out.wav"
title = "GroTTS"
examples = [
['Ze gingen mit klas noar waddendiek, over en deur bragel lopen.', 'gronings']
]
gr.Interface(
inference,
[gr.inputs.Textbox(label="input text", lines=3), gr.inputs.Radio(choices=["gronings", "gronings hoogelandsters", "gronings westerkwartiers", "gronings oldambster"], type="value", default="gronings", label="language")],
[gr.outputs.Audio(type="file", label="Output"), gr.outputs.File()],
title=title,
examples=examples
).launch(enable_queue=True)
|