voice_clone_v2 / TTS /tests /inputs /test_config.json
ahassoun's picture
Upload 3018 files
ee6e328
raw
history blame
1.77 kB
{
"audio":{
"audio_processor": "audio",
"num_mels": 80,
"fft_size": 1024,
"sample_rate": 22050,
"frame_length_ms": null,
"frame_shift_ms": null,
"hop_length": 256,
"win_length": 1024,
"preemphasis": 0.97,
"min_level_db": -100,
"ref_level_db": 20,
"power": 1.5,
"griffin_lim_iters": 30,
"signal_norm": true,
"symmetric_norm": true,
"clip_norm": true,
"max_norm": 4,
"mel_fmin": 0,
"mel_fmax": 8000,
"do_trim_silence": false,
"spec_gain": 20
},
"characters":{
"pad": "_",
"eos": "~",
"bos": "^",
"characters": "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz!'(),-.:;? ",
"punctuations":"!'(),-.:;? ",
"phonemes":"iyɨʉɯuɪʏʊeøɘəɵɤoɛœɜɞʌɔæɐaɶɑɒᵻʘɓǀɗǃʄǂɠǁʛpbtdʈɖcɟkɡqɢʔɴŋɲɳnɱmʙrʀⱱɾɽɸβfvθðszʃʒʂʐçʝxɣχʁħʕhɦɬɮʋɹɻjɰlɭʎʟˈˌːˑʍwɥʜʢʡɕʑɺɧɚ˞ɫʲ"
},
"hidden_size": 128,
"embedding_size": 256,
"text_cleaner": "english_cleaners",
"epochs": 2000,
"lr": 0.003,
"lr_patience": 5,
"lr_decay": 0.5,
"batch_size": 2,
"r": 5,
"mk": 1.0,
"num_loader_workers": 0,
"memory_size": 5,
"save_step": 200,
"data_path": "tests/data/ljspeech/",
"output_path": "result",
"min_seq_len": 0,
"max_seq_len": 300,
"log_dir": "tests/outputs/",
"use_speaker_embedding": false,
"use_gst": true,
"gst": {
"gst_style_input": null,
"gst_use_speaker_embedding": true,
"gst_embedding_dim": 512,
"gst_num_heads": 4,
"gst_num_style_tokens": 10
}
}