| { | |
| "_name_or_path": "models/hf_export/dictabert_large_P1/config.json", | |
| "architectures": [ | |
| "BertForTokenClassification" | |
| ], | |
| "attention_probs_dropout_prob": 0.1, | |
| "classifier_dropout": null, | |
| "hidden_act": "gelu", | |
| "hidden_dropout_prob": 0.1, | |
| "hidden_size": 1024, | |
| "id2label": { | |
| "0": "O", | |
| "1": "B-SUBSTITUTION", | |
| "2": "I-SUBSTITUTION", | |
| "3": "B-QUANTITY", | |
| "4": "I-QUANTITY", | |
| "5": "B-TECHNIQUE", | |
| "6": "I-TECHNIQUE", | |
| "7": "B-ADDITION", | |
| "8": "I-ADDITION" | |
| }, | |
| "initializer_range": 0.02, | |
| "intermediate_size": 4096, | |
| "label2id": { | |
| "B-ADDITION": 7, | |
| "B-QUANTITY": 3, | |
| "B-SUBSTITUTION": 1, | |
| "B-TECHNIQUE": 5, | |
| "I-ADDITION": 8, | |
| "I-QUANTITY": 4, | |
| "I-SUBSTITUTION": 2, | |
| "I-TECHNIQUE": 6, | |
| "O": 0 | |
| }, | |
| "layer_norm_eps": 1e-12, | |
| "max_position_embeddings": 512, | |
| "model_type": "bert", | |
| "newmodern": true, | |
| "num_attention_heads": 16, | |
| "num_hidden_layers": 24, | |
| "pad_token_id": 0, | |
| "position_embedding_type": "absolute", | |
| "torch_dtype": "float32", | |
| "transformers_version": "4.48.0", | |
| "type_vocab_size": 2, | |
| "use_cache": true, | |
| "vocab_size": 128000 | |
| } | |