| mixed_precision: "fp16" | |
| batch_size: 192 | |
| save_interval: 5000 | |
| log_interval: 10 | |
| num_process: 1 # number of GPUs | |
| num_steps: 1000000 | |
| checkpoint_path: "step_1000000.t7" | |
| dataset_params: | |
| # tokenizer: "transfo-xl-wt103" | |
| token_separator: " " # token used for phoneme separator (space) | |
| token_mask: "M" # token used for phoneme mask (M) | |
| # word_separator: 3039 # token used for word separator (<formula>) | |
| # token_maps: "token_maps.pkl" # token map path | |
| max_mel_length: 512 # max phoneme length | |
| word_mask_prob: 0.15 # probability to mask the entire word | |
| phoneme_mask_prob: 0.1 # probability to mask each phoneme | |
| replace_prob: 0.2 # probablity to replace phonemes | |
| model_params: | |
| vocab_size: 69 | |
| hidden_size: 768 | |
| num_attention_heads: 12 | |
| intermediate_size: 2048 | |
| max_position_embeddings: 512 | |
| num_hidden_layers: 12 | |
| dropout: 0.1 |