decoder:
_target_: nemo.collections.tts.modules.tacotron2.Decoder
decoder_rnn_dim: 1024
encoder_embedding_dim: ${model.encoder.encoder_embedding_dim}
gate_threshold: 0.4
max_decoder_steps: 10000
n_frames_per_step: 1 # currently only 1 is supported
n_mel_channels: ${n_mels}
p_attention_dropout: 0.1
p_decoder_dropout: 0.1
prenet_dim: 256
prenet_p_dropout: 0.5
# Attention parameters
attention_dim: 128
attention_rnn_dim: 1024
# AttentionLocation Layer parameters
attention_location_kernel_size: 31
attention_location_n_filters: 32
early_stopping: true
################################################################
! HYDRA_FULL_ERROR=1 \
python tacotron2.py train_dataset=/home/x/2022q2/conv_nemo/manifest/train_tts_6th.json \
validation_datasets=/home/x/2022q2/conv_nemo/manifest/test_tts_6th.json \
trainer.max_epochs=1600 \
trainer.accelerator=null \
trainer.check_val_every_n_epoch=1
谢谢,麻烦了
|