| tokenizer: | |
| type: text_tokenizer.DictTokenizer | |
| args: | |
| max_length: 20 | |
| target_sr: 16000 | |
| model: | |
| args: | |
| shared_dim: 1024 | |
| tchr_dim: 768 | |
| model: | |
| args: {} | |
| decoder: | |
| args: | |
| attn_emb_dim: 1408 | |
| dropout: 0.2 | |
| emb_dim: 256 | |
| fc_emb_dim: 1408 | |
| nlayers: 2 | |
| tie_weights: true | |
| vocab_size: 4981 | |
| type: models.transformer_decoder.TransformerDecoder | |
| encoder: | |
| args: | |
| freeze: false | |
| pretrained: true | |
| type: models.cnn_encoder.EfficientNetB2 | |
| type: models.transformer_model.TransformerModel | |
| type: models.kd_wrapper.ContraEncoderKdWrapper | |