| acous_params: | |
| - - 480 | |
| - 1200 | |
| - 80 | |
| - - 240 | |
| - 1200 | |
| - 160 | |
| amp: false | |
| audio_num_mel_bins: 160 | |
| audio_sample_rate: 24000 | |
| balance_sil: true | |
| c_spk_enc: 512 | |
| char_dict_size: 15000 | |
| conv_use_pos: false | |
| dec0_dilations: | |
| - 1 | |
| - 2 | |
| - 4 | |
| - 1 | |
| - 2 | |
| - 4 | |
| - 1 | |
| dec0_kernel_size: 3 | |
| dec_dilations: | |
| - 1 | |
| - 2 | |
| - 1 | |
| - 2 | |
| - 1 | |
| dec_ffn_kernel_size: 9 | |
| dec_hidden_size: 2048 | |
| dec_inp_add_noise: false | |
| dec_kernel_size: 5 | |
| dec_layers: 4 | |
| dec_num_heads: 8 | |
| dec_post_net_kernel: 3 | |
| decoder_rnn_dim: 0 | |
| decoder_type: conv | |
| dropout: 0.0 | |
| ds_add_pitch_embed: false | |
| dur_alpha: 1.0 | |
| dur_code_size: 128 | |
| dur_context_enc: true | |
| dur_log: true | |
| dur_model_hidden_size: 512 | |
| dur_model_layers: 8 | |
| dur_model_type: ar_mse | |
| dur_predictor_kernel: 3 | |
| dur_predictor_layers: 2 | |
| dur_txt_hs: 512 | |
| dur_use_char: true | |
| dur_use_spk: true | |
| enc_dec_norm: ln | |
| enc_dilations: | |
| - 1 | |
| - 1 | |
| - 1 | |
| - 1 | |
| enc_ffn_kernel_size: 3 | |
| enc_hidden_size: 256 | |
| enc_kernel_size: 5 | |
| enc_layers: 4 | |
| enc_post_net_kernel: 3 | |
| enc_pre_ln: true | |
| enc_prenet: true | |
| encoder_K: 8 | |
| encoder_type: rel_fft | |
| f0_max: 600 | |
| f0_min: 60 | |
| ffn_act: gelu | |
| ffn_hidden_size: 1024 | |
| fft_size: 1200 | |
| fg_spk_enc_hidden: 256 | |
| flatten_dec: true | |
| fmax: 12000 | |
| fmin: 0 | |
| frames_multiple: 8 | |
| hidden_size: 512 | |
| hop_size: 240 | |
| ignore_begin_end_sil: false | |
| lat_for_dur: false | |
| latent_size: 256 | |
| layers_in_block: 2 | |
| ling_label_dict_size: | |
| - 20 | |
| - 4 | |
| - 5 | |
| - 2 | |
| - 3 | |
| - 3 | |
| - 3 | |
| - 6 | |
| - 15 | |
| ling_labels: | |
| - tone | |
| lm_num_layers: 24 | |
| lm_use_enc: true | |
| loud_norm: false | |
| max_tokens: 6000 | |
| mel_vmax: 0.5 | |
| mel_vmin: -6 | |
| min_frames: 0 | |
| mix_melout_timbre: true | |
| mix_ph_timbre: false | |
| model_type: 1 | |
| multistage: false | |
| no_text_enc: false | |
| num_heads: 2 | |
| out_wav_norm: true | |
| pad_frames: false | |
| precision: fp16 | |
| predict_pitch: false | |
| predictor_dropout: 0.0 | |
| predictor_grad: 1.0 | |
| predictor_hidden: -1 | |
| predictor_kernel: 5 | |
| predictor_layers: 5 | |
| print_nan_grads: true | |
| ref_mel_bins: 160 | |
| ref_size_max: 2000 | |
| ref_size_min: 1000 | |
| remove_sil: false | |
| shuffle_ref: false | |
| split_ref: true | |
| temperature: 0.8 | |
| tone_percep_ckpt: '' | |
| train_spk_embed_only: false | |
| use_bert_input: false | |
| use_char: true | |
| use_cur_global: false | |
| use_cur_global_dec: true | |
| use_dur_embed: true | |
| use_dur_mask_embed: true | |
| use_finegrained_spk: false | |
| use_global_lat: false | |
| use_gpt: true | |
| use_gt_dur: false | |
| use_gt_f0: false | |
| use_mix_spk_embed: false | |
| use_new_vae: false | |
| use_ph_level_f0: false | |
| use_ph_pos_embed: true | |
| use_pitch_embed: false | |
| use_pitch_embed_dec: false | |
| use_pitch_pred: true | |
| use_pos_embed: false | |
| use_post_ln: false | |
| use_random_spk_embed: false | |
| use_rot_embed: true | |
| use_spk_embed: false | |
| use_spk_enc: false | |
| use_spk_id: false | |
| use_text_postnet: true | |
| use_uv: true | |
| use_vae: true | |
| use_vqvae: true | |
| use_word_encoder: true | |
| use_word_input: false | |
| vae_dur_grad: 0.1 | |
| vae_enc_hidden_size: 384 | |
| vae_word_conder_layers: 0 | |
| vq_stride: 8 | |
| w_nonsil: 10.0 | |
| w_sil: 1.0 | |
| word_dict_size: 10000 | |
| z_channels: 64 | |
| z_clamp: 2.0 |