Ruth's picture
Training in progress epoch 0
92894c1
{
"_name_or_path": "deepset/gelectra-large",
"architectures": [
"ElectraForTokenClassification"
],
"attention_probs_dropout_prob": 0.1,
"classifier_dropout": null,
"embedding_size": 1024,
"hidden_act": "gelu",
"hidden_dropout_prob": 0.1,
"hidden_size": 1024,
"id2label": {
"0": "O",
"1": "B-LOC",
"10": "I-ORGderiv",
"11": "B-ORGpart",
"12": "I-ORGpart",
"13": "B-OTH",
"14": "I-OTH",
"15": "B-OTHderiv",
"16": "I-OTHderiv",
"17": "B-OTHpart",
"18": "I-OTHpart",
"19": "B-PER",
"2": "I-LOC",
"20": "I-PER",
"21": "B-PERderiv",
"22": "I-PERderiv",
"23": "B-PERpart",
"24": "I-PERpart",
"3": "B-LOCderiv",
"4": "I-LOCderiv",
"5": "B-LOCpart",
"6": "I-LOCpart",
"7": "B-ORG",
"8": "I-ORG",
"9": "B-ORGderiv"
},
"initializer_range": 0.02,
"intermediate_size": 4096,
"label2id": {
"B-LOC": "1",
"B-LOCderiv": "3",
"B-LOCpart": "5",
"B-ORG": "7",
"B-ORGderiv": "9",
"B-ORGpart": "11",
"B-OTH": "13",
"B-OTHderiv": "15",
"B-OTHpart": "17",
"B-PER": "19",
"B-PERderiv": "21",
"B-PERpart": "23",
"I-LOC": "2",
"I-LOCderiv": "4",
"I-LOCpart": "6",
"I-ORG": "8",
"I-ORGderiv": "10",
"I-ORGpart": "12",
"I-OTH": "14",
"I-OTHderiv": "16",
"I-OTHpart": "18",
"I-PER": "20",
"I-PERderiv": "22",
"I-PERpart": "24",
"O": "0"
},
"layer_norm_eps": 1e-12,
"max_position_embeddings": 512,
"model_type": "electra",
"num_attention_heads": 16,
"num_hidden_layers": 24,
"pad_token_id": 0,
"position_embedding_type": "absolute",
"summary_activation": "gelu",
"summary_last_dropout": 0.1,
"summary_type": "first",
"summary_use_proj": true,
"transformers_version": "4.18.0",
"type_vocab_size": 2,
"use_cache": true,
"vocab_size": 31102
}