| { | |
| "_name_or_path": "sberbank-ai/sbert_large_nlu_ru", | |
| "architectures": [ | |
| "BertForSequenceClassification" | |
| ], | |
| "attention_probs_dropout_prob": 0.1, | |
| "classifier_dropout": null, | |
| "directionality": "bidi", | |
| "gradient_checkpointing": false, | |
| "hidden_act": "gelu", | |
| "hidden_dropout_prob": 0.1, | |
| "hidden_size": 1024, | |
| "id2label": { | |
| "0": "ГЛ-000038", | |
| "1": "ГЛ-000039", | |
| "2": "ГЛ-000040", | |
| "3": "ГЛ-000041", | |
| "4": "ГЛ-000045", | |
| "5": "ГЛ-000056", | |
| "6": "ГЛ-000087", | |
| "7": "ГЛ-000088", | |
| "8": "ГЛ-000092", | |
| "9": "ГЛ-000093", | |
| "10": "ГЛ-000095", | |
| "11": "ГЛ-000096", | |
| "12": "ГЛ-000098", | |
| "13": "ГЛ-000099", | |
| "14": "ГЛ-000101", | |
| "15": "ГЛ-000104", | |
| "16": "ГЛ-000105", | |
| "17": "ГЛ-000106", | |
| "18": "ГЛ-000107", | |
| "19": "ГЛ-000108", | |
| "20": "ГЛ-000109", | |
| "21": "ГЛ-000114", | |
| "22": "ГЛ-000115", | |
| "23": "ГЛ-000123", | |
| "24": "ГЛ-000124", | |
| "25": "ГЛ-000130", | |
| "26": "ГЛ-000132" | |
| }, | |
| "initializer_range": 0.02, | |
| "intermediate_size": 4096, | |
| "label2id": { | |
| "ГЛ-000038": 0, | |
| "ГЛ-000039": 1, | |
| "ГЛ-000095": 10, | |
| "ГЛ-000096": 11, | |
| "ГЛ-000098": 12, | |
| "ГЛ-000099": 13, | |
| "ГЛ-000101": 14, | |
| "ГЛ-000104": 15, | |
| "ГЛ-000105": 16, | |
| "ГЛ-000106": 17, | |
| "ГЛ-000107": 18, | |
| "ГЛ-000108": 19, | |
| "ГЛ-000040": 2, | |
| "ГЛ-000109": 20, | |
| "ГЛ-000114": 21, | |
| "ГЛ-000115": 22, | |
| "ГЛ-000123": 23, | |
| "ГЛ-000124": 24, | |
| "ГЛ-000130": 25, | |
| "ГЛ-000132": 26, | |
| "ГЛ-000041": 3, | |
| "ГЛ-000045": 4, | |
| "ГЛ-000056": 5, | |
| "ГЛ-000087": 6, | |
| "ГЛ-000088": 7, | |
| "ГЛ-000092": 8, | |
| "ГЛ-000093": 9 | |
| }, | |
| "layer_norm_eps": 1e-12, | |
| "max_position_embeddings": 512, | |
| "model_type": "bert", | |
| "num_attention_heads": 16, | |
| "num_hidden_layers": 24, | |
| "pad_token_id": 0, | |
| "pooler_fc_size": 768, | |
| "pooler_num_attention_heads": 12, | |
| "pooler_num_fc_layers": 3, | |
| "pooler_size_per_head": 128, | |
| "pooler_type": "first_token_transform", | |
| "position_embedding_type": "absolute", | |
| "problem_type": "single_label_classification", | |
| "torch_dtype": "float32", | |
| "transformers_version": "4.19.2", | |
| "type_vocab_size": 2, | |
| "use_cache": true, | |
| "vocab_size": 120138 | |
| } | |