| | |
|---|
| | | # tables.print() |
|---|
| | | |
|---|
| | | # network architecture |
|---|
| | | #model: funasr.models.paraformer.model:Paraformer |
|---|
| | | model: Transformer |
|---|
| | | model: Conformer |
|---|
| | | model_conf: |
|---|
| | | ctc_weight: 0.3 |
|---|
| | | lsm_weight: 0.1 # label smoothing option |
|---|
| | |
|---|
| | | # encoder |
|---|
| | | encoder: ConformerEncoder |
|---|
| | | encoder_conf: |
|---|
| | | output_size: 256 # dimension of attention |
|---|
| | | output_size: 256 |
|---|
| | | attention_heads: 4 |
|---|
| | | linear_units: 2048 # the number of units of position-wise feed forward |
|---|
| | | num_blocks: 12 # the number of encoder blocks |
|---|
| | | linear_units: 2048 |
|---|
| | | num_blocks: 12 |
|---|
| | | dropout_rate: 0.1 |
|---|
| | | positional_dropout_rate: 0.1 |
|---|
| | | attention_dropout_rate: 0.0 |
|---|
| | | input_layer: conv2d # encoder architecture type |
|---|
| | | input_layer: conv2d |
|---|
| | | normalize_before: true |
|---|
| | | pos_enc_layer_type: rel_pos |
|---|
| | | selfattention_layer_type: rel_selfattn |
|---|
| | |
|---|
| | | n_mels: 80 |
|---|
| | | frame_length: 25 |
|---|
| | | frame_shift: 10 |
|---|
| | | dither: 0.0 |
|---|
| | | lfr_m: 1 |
|---|
| | | lfr_n: 1 |
|---|
| | | |
|---|
| | |
|---|
| | | dataset: AudioDataset |
|---|
| | | dataset_conf: |
|---|
| | | index_ds: IndexDSJsonl |
|---|
| | | batch_sampler: DynamicBatchLocalShuffleSampler |
|---|
| | | batch_sampler: BatchSampler |
|---|
| | | batch_type: example # example or length |
|---|
| | | batch_size: 1 # 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, |
|---|
