RMSnow's picture
add backend inference and inferface output
0883aa1
{
"base_config": "config/base.json",
"supported_model_type": [
"Fastspeech2",
"VITS",
"VALLE",
],
"task_type": "tts",
"preprocess": {
"language": "en-us",
// linguistic features
"extract_phone": true,
"phone_extractor": "espeak", // "espeak, pypinyin, pypinyin_initials_finals, lexicon (only for language=en-us right now)"
"lexicon_path": "./text/lexicon/librispeech-lexicon.txt",
// Directory names of processed data or extracted features
"phone_dir": "phones",
"use_phone": true,
},
"model": {
"text_token_num": 512,
}
}