Text-to-Speech
PyTorch
ONNX
Catalan
matcha-tts
acoustic modelling
speech
multispeaker
AlexK-PL's picture
Upload model and config files
2842206 verified
raw
history blame contribute delete
736 Bytes
cfm:
name: CFM
sigma_min: 0.0001
solver: euler
data_statistics:
mel_mean: -6.578195
mel_std: 2.538758
decoder:
act_fn: snakebeta
attention_head_dim: 64
channels:
- 256
- 256
dropout: 0.05
n_blocks: 1
num_heads: 2
num_mid_blocks: 2
encoder:
duration_predictor_params:
filter_channels_dp: 256
kernel_size: 3
p_dropout: 0.1
encoder_params:
filter_channels: 768
filter_channels_dp: 256
kernel_size: 3
n_channels: 192
n_feats: 80
n_heads: 2
n_layers: 6
n_spks: 47
p_dropout: 0.1
prenet: true
spk_emb_dim: 64
encoder_type: RoPE Encoder
n_feats: 80
n_spks: 47
n_vocab: 178
optimizer: null
out_size: null
prior_loss: true
scheduler: null
spk_emb_dim: 64