initial_model / config.json
gykim80's picture
Training in progress, epoch 1
c28c91c verified
{
"_name_or_path": "monologg/kobert",
"architectures": [
"BertForSequenceClassification"
],
"attention_probs_dropout_prob": 0.1,
"classifier_dropout": null,
"hidden_act": "gelu",
"hidden_dropout_prob": 0.1,
"hidden_size": 768,
"id2label": {
"0": "55",
"1": "92",
"2": "73",
"3": "67",
"4": "28",
"5": "61",
"6": "34",
"7": "49",
"8": "81",
"9": "20",
"10": "07",
"11": "74",
"12": "62",
"13": "63",
"14": "14",
"15": "25",
"16": "56",
"17": "75",
"18": "66",
"19": "52",
"20": "21",
"21": "05",
"22": "03",
"23": "68",
"24": "70",
"25": "38",
"26": "51",
"27": "65",
"28": "76",
"29": "57",
"30": "35",
"31": "60",
"32": "94",
"33": "27",
"34": "87",
"35": "26",
"36": "72",
"37": "37",
"38": "32",
"39": "85",
"40": "95",
"41": "29",
"42": "33",
"43": "71",
"44": "47",
"45": "90",
"46": "42",
"47": "58",
"48": "84",
"49": "83",
"50": "96",
"51": "97",
"52": "44",
"53": "48",
"54": "08",
"55": "54",
"56": "64",
"57": "86",
"58": "39",
"59": "93",
"60": "59",
"61": "13",
"62": "23",
"63": "41",
"64": "88",
"65": "80",
"66": "40",
"67": "69",
"68": "82",
"69": "31"
},
"initializer_range": 0.02,
"intermediate_size": 3072,
"label2id": {
"03": 22,
"05": 21,
"07": 10,
"08": 54,
"13": 61,
"14": 14,
"20": 9,
"21": 20,
"23": 62,
"25": 15,
"26": 35,
"27": 33,
"28": 4,
"29": 41,
"31": 69,
"32": 38,
"33": 42,
"34": 6,
"35": 30,
"37": 37,
"38": 25,
"39": 58,
"40": 66,
"41": 63,
"42": 46,
"44": 52,
"47": 44,
"48": 53,
"49": 7,
"51": 26,
"52": 19,
"54": 55,
"55": 0,
"56": 16,
"57": 29,
"58": 47,
"59": 60,
"60": 31,
"61": 5,
"62": 12,
"63": 13,
"64": 56,
"65": 27,
"66": 18,
"67": 3,
"68": 23,
"69": 67,
"70": 24,
"71": 43,
"72": 36,
"73": 2,
"74": 11,
"75": 17,
"76": 28,
"80": 65,
"81": 8,
"82": 68,
"83": 49,
"84": 48,
"85": 39,
"86": 57,
"87": 34,
"88": 64,
"90": 45,
"92": 1,
"93": 59,
"94": 32,
"95": 40,
"96": 50,
"97": 51
},
"layer_norm_eps": 1e-12,
"max_position_embeddings": 512,
"model_type": "bert",
"num_attention_heads": 12,
"num_hidden_layers": 12,
"pad_token_id": 1,
"position_embedding_type": "absolute",
"problem_type": "single_label_classification",
"torch_dtype": "float32",
"transformers_version": "4.48.2",
"type_vocab_size": 2,
"use_cache": true,
"vocab_size": 8002
}