| log_dir: "Models/Anispeech_with_DIFF" | |
| first_stage_path: "/home/ubuntu/StyleTTS_Accelerate_44khz/Models/Anispeech/epoch_1st_00020.pth" | |
| save_freq: 1 | |
| log_interval: 10 | |
| device: "cuda" | |
| multigpu: false | |
| epochs_1st: 200 # number of epochs for first stage training | |
| epochs_2nd: 100 # number of peochs for second stage training | |
| batch_size: 32 | |
| pretrained_model: "" | |
| second_stage_load_pretrained: true # set to true if the pre-trained model is for 2nd stage | |
| load_only_params: false # set to true if do not want to load epoch numbers and optimizer parameters | |
| diff_epoch: 5 | |
| train_data: "/home/ubuntu/StyleTTS_Accelerate_44khz/Data/ani_train_only_longs.csv" | |
| val_data: "/home/ubuntu/StyleTTS_Accelerate_44khz/Data/val_list_libritts.txt" | |
| F0_path: "Utils/JDC/bst.t7" | |
| ASR_config: "Utils/ASR/config.yml" | |
| ASR_path: "Utils/ASR/epoch_00080.pth" | |
| preprocess_params: | |
| sr: 24000 | |
| spect_params: | |
| n_fft: 2048 | |
| win_length: 1200 | |
| hop_length: 300 | |
| model_params: | |
| hidden_dim: 512 | |
| n_token: 178 | |
| style_dim: 128 | |
| n_layer: 3 | |
| dim_in: 64 | |
| max_conv_dim: 512 | |
| n_mels: 80 | |
| dropout: 0.2 | |
| diffusion: | |
| embedding_mask_proba: 0.1 | |
| # transformer config | |
| transformer: | |
| num_layers: 3 | |
| num_heads: 8 | |
| head_features: 64 | |
| multiplier: 2 | |
| # diffusion distribution config | |
| dist: | |
| sigma_data: 0.2 # placeholder for estimate_sigma_data set to false | |
| estimate_sigma_data: true # estimate sigma_data from the current batch if set to true | |
| mean: -3.0 | |
| std: 1.0 | |
| loss_params: | |
| lambda_mel: 10. # mel reconstruction loss (1st & 2nd stage) | |
| lambda_adv: 1. # adversarial loss (1st & 2nd stage) | |
| lambda_reg: 1. # adversarial regularization loss (1st & 2nd stage) | |
| lambda_fm: 0.1 # feature matching loss (1st & 2nd stage) | |
| lambda_mono: 1. # monotonic alignment loss (1st stage, TMA) | |
| lambda_s2s: 1. # sequence-to-sequence loss (1st stage, TMA) | |
| TMA_epoch: 2 # TMA starting epoch (1st stage) | |
| # https://github.com/yl4579/StyleTTS/issues/7 | |
| TMA_CEloss: false # whether to use cross-entropy (CE) loss for TMA | |
| lambda_F0: 1. # F0 reconstruction loss (2nd stage) | |
| lambda_norm: 1. # norm reconstruction loss (2nd stage) | |
| lambda_dur: 1. # duration loss (2nd stage) | |
| lambda_ce: 20. # duration predictor probability output CE loss (2nd stage) | |
| lambda_sty: 1. # style reconstruction loss (2nd stage) | |
| lambda_diff: 1. # score matching loss (2nd stage) | |
| optimizer_params: | |
| lr: 0.0001 | |