histbert-finetuned-ner / config.json
crina-t's picture
Training in progress, epoch 1
719ccdc
raw
history blame
1.32 kB
{
"_name_or_path": "Riksarkivet/bert-base-cased-swe-historical",
"architectures": [
"BertForTokenClassification"
],
"attention_probs_dropout_prob": 0.1,
"classifier_dropout": null,
"gradient_checkpointing": false,
"hidden_act": "gelu",
"hidden_dropout_prob": 0.1,
"hidden_size": 768,
"id2label": {
"0": "O",
"1": "B-EVN",
"2": "B-LOC",
"3": "B-MSR",
"4": "B-OBJ",
"5": "B-ORG",
"6": "B-PRS",
"7": "B-TME",
"8": "B-WRK",
"9": "I-EVN",
"10": "I-LOC",
"11": "I-MSR",
"12": "I-OBJ",
"13": "I-ORG",
"14": "I-PRS",
"15": "I-TME",
"16": "I-WRK"
},
"initializer_range": 0.02,
"intermediate_size": 3072,
"label2id": {
"B-EVN": 1,
"B-LOC": 2,
"B-MSR": 3,
"B-OBJ": 4,
"B-ORG": 5,
"B-PRS": 6,
"B-TME": 7,
"B-WRK": 8,
"I-EVN": 9,
"I-LOC": 10,
"I-MSR": 11,
"I-OBJ": 12,
"I-ORG": 13,
"I-PRS": 14,
"I-TME": 15,
"I-WRK": 16,
"O": 0
},
"layer_norm_eps": 1e-12,
"max_position_embeddings": 512,
"model_type": "bert",
"num_attention_heads": 12,
"num_hidden_layers": 12,
"pad_token_id": 0,
"position_embedding_type": "absolute",
"torch_dtype": "float32",
"transformers_version": "4.28.0",
"type_vocab_size": 2,
"use_cache": true,
"vocab_size": 64000
}