fias_NER / config.json
Ventalik's picture
Training in progress, epoch 1
dff4747
{
"_name_or_path": "cointegrated/rubert-tiny2",
"architectures": [
"BertForTokenClassification"
],
"attention_probs_dropout_prob": 0.1,
"classifier_dropout": null,
"emb_size": 312,
"gradient_checkpointing": false,
"hidden_act": "gelu",
"hidden_dropout_prob": 0.1,
"hidden_size": 312,
"id2label": {
"0": "Out",
"1": "\u0421\u0443\u0431\u044a\u0435\u043a\u0442_\u0420\u0424",
"2": "\u0410\u0434\u043c\u0438\u043d\u0438\u0441\u0442\u0440\u0430\u0442\u0438\u0432\u043d\u044b\u0439_\u0440\u0430\u0439\u043e\u043d",
"3": "\u041c\u0443\u043d\u0438\u0446\u0438\u043f\u0430\u043b\u044c\u043d\u044b\u0439_\u0440\u0430\u0439\u043e\u043d",
"4": "\u0421\u0435\u043b\u044c\u0441\u043a\u043e\u0435_\u0433\u043e\u0440\u043e\u0434\u0441\u043a\u043e\u0435_\u043f\u043e\u0441\u0435\u043b\u0435\u043d\u0438\u0435",
"5": "\u0413\u043e\u0440\u043e\u0434",
"6": "\u041d\u0430\u0441\u0435\u043b\u0435\u043d\u043d\u044b\u0439_\u043f\u0443\u043d\u043a\u0442",
"7": "\u042d\u043b\u0435\u043c\u0435\u043d\u0442_\u043f\u043b\u0430\u043d\u0438\u0440\u043e\u0432\u043e\u0447\u043d\u043e\u0439_\u0441\u0442\u0440\u0443\u043a\u0442\u0443\u0440\u044b",
"8": "\u042d\u043b\u0435\u043c\u0435\u043d\u0442_\u0443\u043b\u0438\u0447\u043d\u043e\u0439_\u0441\u0435\u0442\u0438",
"9": "\u0417\u0435\u043c\u0435\u043b\u044c\u043d\u044b\u0439_\u0443\u0447\u0430\u0441\u0442\u043e\u043a",
"10": "\u0417\u0434\u0430\u043d\u0438\u0435",
"11": "\u041f\u043e\u0447\u0442\u043e\u0432\u044b\u0439_\u0438\u043d\u0434\u0435\u043a\u0441"
},
"initializer_range": 0.02,
"intermediate_size": 600,
"label2id": {
"Out": 0,
"\u0410\u0434\u043c\u0438\u043d\u0438\u0441\u0442\u0440\u0430\u0442\u0438\u0432\u043d\u044b\u0439_\u0440\u0430\u0439\u043e\u043d": 2,
"\u0413\u043e\u0440\u043e\u0434": 5,
"\u0417\u0434\u0430\u043d\u0438\u0435": 10,
"\u0417\u0435\u043c\u0435\u043b\u044c\u043d\u044b\u0439_\u0443\u0447\u0430\u0441\u0442\u043e\u043a": 9,
"\u041c\u0443\u043d\u0438\u0446\u0438\u043f\u0430\u043b\u044c\u043d\u044b\u0439_\u0440\u0430\u0439\u043e\u043d": 3,
"\u041d\u0430\u0441\u0435\u043b\u0435\u043d\u043d\u044b\u0439_\u043f\u0443\u043d\u043a\u0442": 6,
"\u041f\u043e\u0447\u0442\u043e\u0432\u044b\u0439_\u0438\u043d\u0434\u0435\u043a\u0441": 11,
"\u0421\u0435\u043b\u044c\u0441\u043a\u043e\u0435_\u0433\u043e\u0440\u043e\u0434\u0441\u043a\u043e\u0435_\u043f\u043e\u0441\u0435\u043b\u0435\u043d\u0438\u0435": 4,
"\u0421\u0443\u0431\u044a\u0435\u043a\u0442_\u0420\u0424": 1,
"\u042d\u043b\u0435\u043c\u0435\u043d\u0442_\u043f\u043b\u0430\u043d\u0438\u0440\u043e\u0432\u043e\u0447\u043d\u043e\u0439_\u0441\u0442\u0440\u0443\u043a\u0442\u0443\u0440\u044b": 7,
"\u042d\u043b\u0435\u043c\u0435\u043d\u0442_\u0443\u043b\u0438\u0447\u043d\u043e\u0439_\u0441\u0435\u0442\u0438": 8
},
"layer_norm_eps": 1e-12,
"max_position_embeddings": 2048,
"model_type": "bert",
"num_attention_heads": 12,
"num_hidden_layers": 3,
"pad_token_id": 0,
"position_embedding_type": "absolute",
"torch_dtype": "float32",
"transformers_version": "4.30.2",
"type_vocab_size": 2,
"use_cache": true,
"vocab_size": 83828
}