표 1. | Table 1. 실험에 사용한 음향 모델의 하이퍼파라미터 | Hyperparameters of acoustic model

Hyperparameter	Value
Phoneme embedding dimension	256
Encoder layers	4
Encoder hidden	256
Encoder Conv1D kernel	9
Encoder Conv1D filter size	1,024
Encoder attention heads	2
Mel-spectrogram decoder layers	4
Mel-spectrogram decoder hidden	256
Mel-spectrogram decoder Conv1D kernel	9
Mel-spectrogram decoder Conv1D filter size	1,024
Mel-spectrogram decoder attention headers	2
Encoder / decoder dropout	0.1
Variance predictor Conv1D kernel	3
Variance predictor Conv1D filter size	256
Variance predictor dropout	0.5
Waveform decoder convolution blocks	30
Waveform decoder dilated Conv1D kernel size	3
Waveform decoder transposed Conv1D filter size	64
Waveform decoder skip channel Size	64
Batch size	32