Text-to-Speech
WaveGen
Safetensors
PyTorch
English
Chinese
WaveGen
TTS
zero-shot
voice-cloning
audio
multilingual
Instructions to use LeeSangHoon/WaveGen with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- WaveGen
How to use LeeSangHoon/WaveGen with WaveGen:
# No code snippets available yet for this library. # To use this model, check the repository files and the library's documentation. # Want to help? PRs adding snippets are welcome at: # https://github.com/huggingface/huggingface.js
- Notebooks
- Google Colab
- Kaggle
File size: 1,223 Bytes
5d9f2da | 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 | model:
name: WaveGen_Emilia_24k_50Hz_SF_80bin_ibdim32_drop02
tokenizer: pinyin
backbone: DiT
cfm:
prediction: x_pred
loss_space: v
t_sampling: logistic_normal
P_mean: -0.8
P_std: 0.8
time_shift: 1.0
t_eps: 0.02
latents_scale: 9.0
use_aux_mel_loss: false
aux_mel_loss_weight: 0.0
use_ms_v_spec_loss: true
ms_v_spec_loss_weight: 0.1
aux_mel_loss_masked: true
sf_loss_weight: 0.1
mel_xt_drop_prob: 0.0
inference_bottleneck_dropout: 0.2
inference_bottleneck_dropout_seed: 0
sample_rate: 24000
joint_cond_drop_prob: 0.1
arch:
dim: 1152
depth: 24
heads: 16
dropout: 0.0
ff_mult: 4
text_dim: 512
text_mask_padding: true
qk_norm: null
conv_layers: 4
pe_attn_head: null
attn_backend: torch
attn_mask_enabled: false
checkpoint_activations: false
mel_channels: 80
use_mel_information_bottleneck: true
mel_bottleneck_dim: 32
mel_bottleneck_dropout: 0.2
mel_bottleneck_zero_init: false
use_audio_proj: true
audio_proj_dim: 1024
audio_proj_hidden: 2048
waveform:
target_sample_rate: 24000
wav_frame_len: 480
|