cedricbonhomme's picture
End of training
40ad73e verified
Raw
History Blame Contribute Delete
2.81 kB
{
"add_cross_attention": false,
"architectures": [
"RobertaForSequenceClassification"
],
"attention_probs_dropout_prob": 0.1,
"bos_token_id": 0,
"classifier_dropout": null,
"dtype": "float32",
"eos_token_id": 2,
"hidden_act": "gelu",
"hidden_dropout_prob": 0.1,
"hidden_size": 768,
"id2label": {
"0": "T1003",
"1": "T1005",
"2": "T1021",
"3": "T1027",
"4": "T1036",
"5": "T1040",
"6": "T1041",
"7": "T1046",
"8": "T1055",
"9": "T1059",
"10": "T1068",
"11": "T1070",
"12": "T1071",
"13": "T1078",
"14": "T1082",
"15": "T1083",
"16": "T1087",
"17": "T1091",
"18": "T1098",
"19": "T1105",
"20": "T1106",
"21": "T1110",
"22": "T1114",
"23": "T1133",
"24": "T1136",
"25": "T1185",
"26": "T1189",
"27": "T1190",
"28": "T1202",
"29": "T1203",
"30": "T1204",
"31": "T1210",
"32": "T1211",
"33": "T1212",
"34": "T1485",
"35": "T1486",
"36": "T1489",
"37": "T1496",
"38": "T1497",
"39": "T1498",
"40": "T1499",
"41": "T1505",
"42": "T1528",
"43": "T1542",
"44": "T1543",
"45": "T1548",
"46": "T1550",
"47": "T1552",
"48": "T1555",
"49": "T1557",
"50": "T1563",
"51": "T1565",
"52": "T1566",
"53": "T1574",
"54": "T1588",
"55": "T1608",
"56": "T1685"
},
"initializer_range": 0.02,
"intermediate_size": 3072,
"is_decoder": false,
"label2id": {
"T1003": 0,
"T1005": 1,
"T1021": 2,
"T1027": 3,
"T1036": 4,
"T1040": 5,
"T1041": 6,
"T1046": 7,
"T1055": 8,
"T1059": 9,
"T1068": 10,
"T1070": 11,
"T1071": 12,
"T1078": 13,
"T1082": 14,
"T1083": 15,
"T1087": 16,
"T1091": 17,
"T1098": 18,
"T1105": 19,
"T1106": 20,
"T1110": 21,
"T1114": 22,
"T1133": 23,
"T1136": 24,
"T1185": 25,
"T1189": 26,
"T1190": 27,
"T1202": 28,
"T1203": 29,
"T1204": 30,
"T1210": 31,
"T1211": 32,
"T1212": 33,
"T1485": 34,
"T1486": 35,
"T1489": 36,
"T1496": 37,
"T1497": 38,
"T1498": 39,
"T1499": 40,
"T1505": 41,
"T1528": 42,
"T1542": 43,
"T1543": 44,
"T1548": 45,
"T1550": 46,
"T1552": 47,
"T1555": 48,
"T1557": 49,
"T1563": 50,
"T1565": 51,
"T1566": 52,
"T1574": 53,
"T1588": 54,
"T1608": 55,
"T1685": 56
},
"layer_norm_eps": 1e-05,
"max_position_embeddings": 514,
"model_type": "roberta",
"num_attention_heads": 12,
"num_hidden_layers": 12,
"pad_token_id": 1,
"problem_type": "multi_label_classification",
"tie_word_embeddings": true,
"transformers_version": "5.13.0",
"type_vocab_size": 1,
"use_cache": false,
"vocab_size": 50265
}