arch: pdss2t_transformer_m_8

encoder-embed-dim: 512
pds-stages: 4
pds-layers: 3_3_3_3
pds-ratios: 2_2_1_2
pds-fusion: True
pds-fusion-method: all_conv
pds-embed-dims: 512_512_512_512
pds-ds-method: conv
pds-embed-norm: True
pds-position-embed: 1_1_1_1
pds-kernel-sizes: 5_5_5_5
pds-ffn-ratios: 4_4_4_4
pds-attn-heads: 8_8_8_8

share-decoder-input-output-embed: True
optimizer: adam
clip-norm: 10.0
lr-scheduler: inverse_sqrt
warmup-init-lr: 1e-7
warmup-updates: 10000
lr: 0.0014
adam_betas: (0.9,0.98)

criterion: label_smoothed_cross_entropy_with_ctc
label_smoothing: 0.1

dropout: 0.15
activation-fn: relu
encoder-ffn-embed-dim: 2048
encoder-layers: 12
decoder-layers: 6
encoder-attention-heads: 8

decoder-embed-dim: 512
decoder-ffn-embed-dim: 2048
decoder-attention-heads: 8