Text-to-Speech
Safetensors
cutetts
voice-cloning
CuteTTS / weights /audio_vae /config.json
MinMinLiang's picture
Upload 16 files
5104922 verified
Raw
History Blame Contribute Delete
466 Bytes
{
"component": "audio_vae",
"channels": 1,
"decoder_dim": 1536,
"decoder_rates": [
16,
8,
5,
3
],
"depthwise": true,
"encoder_dim": 128,
"encoder_rates": [
3,
5,
8,
16
],
"fix_std": 0.15,
"frame_rate": 12.5,
"logvar_max": 20.0,
"logvar_min": -30.0,
"posterior_type": "sigma",
"sample_rate": 24000,
"std_dist_type": "gaussian",
"use_noise_block": false,
"vae_dim": 64,
"weight_init_value": 0.02
}