| | |
|---|
| | | accum_grad: 1 |
|---|
| | | grad_clip: 5 |
|---|
| | | max_epoch: 150 |
|---|
| | | val_scheduler_criterion: |
|---|
| | | - valid |
|---|
| | | - acc |
|---|
| | | best_model_criterion: |
|---|
| | | - - valid |
|---|
| | | - acc |
|---|
| | | - max |
|---|
| | | keep_nbest_models: 10 |
|---|
| | | log_interval: 50 |
|---|
| | | |
|---|
| | |
|---|
| | | dataset: AudioDataset |
|---|
| | | dataset_conf: |
|---|
| | | index_ds: IndexDSJsonl |
|---|
| | | batch_sampler: RankFullLocalShuffleBatchSampler |
|---|
| | | batch_type: example # example or length |
|---|
| | | batch_size: 32 # if batch_type is example, batch_size is the numbers of samples; if length, batch_size is source_token_len+target_token_len; |
|---|
| | | batch_sampler: EspnetStyleBatchSampler |
|---|
| | | batch_type: length # example or length |
|---|
| | | batch_size: 25000 # if batch_type is example, batch_size is the numbers of samples; if length, batch_size is source_token_len+target_token_len; |
|---|
| | | max_token_length: 2048 # filter samples if source_token_len+target_token_len > max_token_length, |
|---|
| | | buffer_size: 1024 |
|---|
| | | shuffle: True |
|---|
| | |
|---|
| | | tokenizer: CharTokenizer |
|---|
| | | tokenizer_conf: |
|---|
| | | unk_symbol: <unk> |
|---|
| | | split_with_space: true |
|---|
| | | |
|---|
| | | |
|---|
| | | ctc_conf: |
|---|
| | | dropout_rate: 0.0 |
|---|
