| name: roberta_sequence_labeling | |
| config_type: model | |
| task: sequence_labeling | |
| num_labels: 23 | |
| id2label: | |
| 0: P | |
| 1: Ne | |
| 2: PRO | |
| 3: CONJ | |
| 4: 'N' | |
| 5: PUNC | |
| 6: Pe | |
| 7: ADV | |
| 8: V | |
| 9: AJ | |
| 10: AJe | |
| 11: DET | |
| 12: POSTP | |
| 13: NUM | |
| 14: DETe | |
| 15: NUMe | |
| 16: PROe | |
| 17: ADVe | |
| 18: RES | |
| 19: CL | |
| 20: INT | |
| 21: CONJe | |
| 22: RESe | |
| attention_probs_dropout_prob: 0.1 | |
| bos_token_id: 0 | |
| eos_token_id: 2 | |
| gradient_checkpointing: false | |
| hidden_act: gelu | |
| hidden_dropout_prob: 0.1 | |
| hidden_size: 768 | |
| initializer_range: 0.02 | |
| intermediate_size: 3072 | |
| layer_norm_eps: 1.0e-12 | |
| max_position_embeddings: 514 | |
| num_attention_heads: 12 | |
| num_hidden_layers: 12 | |
| pad_token_id: 1 | |
| position_embedding_type: absolute | |
| type_vocab_size: 1 | |
| use_cache: true | |
| vocab_size: 42000 | |
| prediction_skip_tokens: | |
| - <s> | |
| - <pad> | |
| - </s> | |
| - <unk> | |
| - <mask> | |
| - <|endoftext|> | |
| - <|startoftext|> | |
| - <nl> | |
| - <hs> | |
| - <sep> | |
| - <cls> | |