Training in progress, step 500

.gitignore

@@ -0,0 +1 @@
+checkpoint-*/

config.json

@@ -0,0 +1,260 @@
+{
+  "_name_or_path": "distilbert-base-uncased",
+  "activation": "gelu",
+  "architectures": [
+    "DistilBertForTokenClassification"
+  ],
+  "attention_dropout": 0.1,
+  "dim": 768,
+  "dropout": 0.1,
+  "hidden_dim": 3072,
+  "id2label": {
+    "0": "B-PREFIX",
+    "1": "I-PREFIX",
+    "2": "B-FIRSTNAME",
+    "3": "I-FIRSTNAME",
+    "4": "B-MIDDLENAME",
+    "5": "B-LASTNAME",
+    "6": "I-LASTNAME",
+    "7": "O",
+    "8": "B-JOBDESCRIPTOR",
+    "9": "B-JOBTITLE",
+    "10": "I-JOBTITLE",
+    "11": "B-COMPANY_NAME",
+    "12": "I-COMPANY_NAME",
+    "13": "B-JOBAREA",
+    "14": "B-EMAIL",
+    "15": "I-EMAIL",
+    "16": "B-TIME",
+    "17": "I-TIME",
+    "18": "B-DATE",
+    "19": "I-DATE",
+    "20": "B-URL",
+    "21": "I-URL",
+    "22": "B-BITCOINADDRESS",
+    "23": "I-BITCOINADDRESS",
+    "24": "B-ETHEREUMADDRESS",
+    "25": "I-ETHEREUMADDRESS",
+    "26": "B-ACCOUNTNAME",
+    "27": "I-ACCOUNTNAME",
+    "28": "B-IBAN",
+    "29": "I-IBAN",
+    "30": "B-ACCOUNTNUMBER",
+    "31": "I-ACCOUNTNUMBER",
+    "32": "B-BIC",
+    "33": "I-BIC",
+    "34": "B-IPV4",
+    "35": "I-IPV4",
+    "36": "B-STREETADDRESS",
+    "37": "I-STREETADDRESS",
+    "38": "B-CITY",
+    "39": "I-CITY",
+    "40": "B-ZIPCODE",
+    "41": "I-ZIPCODE",
+    "42": "B-USERNAME",
+    "43": "I-USERNAME",
+    "44": "B-IPV6",
+    "45": "I-IPV6",
+    "46": "B-CREDITCARDNUMBER",
+    "47": "I-CREDITCARDNUMBER",
+    "48": "B-VEHICLEVIN",
+    "49": "I-VEHICLEVIN",
+    "50": "B-SUFFIX",
+    "51": "I-SUFFIX",
+    "52": "B-AMOUNT",
+    "53": "I-AMOUNT",
+    "54": "B-CURRENCY",
+    "55": "I-CURRENCY",
+    "56": "B-PASSWORD",
+    "57": "I-PASSWORD",
+    "58": "B-JOBTYPE",
+    "59": "B-STATE",
+    "60": "B-BUILDINGNUMBER",
+    "61": "I-BUILDINGNUMBER",
+    "62": "B-VEHICLEVRM",
+    "63": "I-VEHICLEVRM",
+    "64": "B-PHONEIMEI",
+    "65": "I-PHONEIMEI",
+    "66": "I-JOBAREA",
+    "67": "I-STATE",
+    "68": "B-COUNTY",
+    "69": "B-CURRENCYNAME",
+    "70": "I-CURRENCYNAME",
+    "71": "B-CURRENCYSYMBOL",
+    "72": "B-MASKEDNUMBER",
+    "73": "I-MASKEDNUMBER",
+    "74": "B-PHONE_NUMBER",
+    "75": "I-PHONE_NUMBER",
+    "76": "B-SECONDARYADDRESS",
+    "77": "I-SECONDARYADDRESS",
+    "78": "B-SSN",
+    "79": "I-SSN",
+    "80": "B-CURRENCYCODE",
+    "81": "B-LITECOINADDRESS",
+    "82": "I-LITECOINADDRESS",
+    "83": "B-MAC",
+    "84": "I-MAC",
+    "85": "B-CREDITCARDISSUER",
+    "86": "I-CREDITCARDISSUER",
+    "87": "B-CREDITCARDCVV",
+    "88": "I-CREDITCARDCVV",
+    "89": "B-USERAGENT",
+    "90": "I-USERAGENT",
+    "91": "B-IP",
+    "92": "I-IP",
+    "93": "B-SEX",
+    "94": "B-STREET",
+    "95": "I-STREET",
+    "96": "B-PIN",
+    "97": "I-PIN",
+    "98": "I-JOBTYPE",
+    "99": "I-MIDDLENAME",
+    "100": "I-CURRENCYCODE",
+    "101": "I-CURRENCYSYMBOL",
+    "102": "B-FULLNAME",
+    "103": "I-FULLNAME",
+    "104": "B-NAME",
+    "105": "I-NAME",
+    "106": "B-GENDER",
+    "107": "B-NUMBER",
+    "108": "I-NUMBER",
+    "109": "I-GENDER",
+    "110": "B-NEARBYGPSCOORDINATE",
+    "111": "I-NEARBYGPSCOORDINATE",
+    "112": "B-DISPLAYNAME",
+    "113": "I-DISPLAYNAME",
+    "114": "B-SEXTYPE",
+    "115": "B-ORDINALDIRECTION"
+  },
+  "initializer_range": 0.02,
+  "label2id": {
+    "B-ACCOUNTNAME": 26,
+    "B-ACCOUNTNUMBER": 30,
+    "B-AMOUNT": 52,
+    "B-BIC": 32,
+    "B-BITCOINADDRESS": 22,
+    "B-BUILDINGNUMBER": 60,
+    "B-CITY": 38,
+    "B-COMPANY_NAME": 11,
+    "B-COUNTY": 68,
+    "B-CREDITCARDCVV": 87,
+    "B-CREDITCARDISSUER": 85,
+    "B-CREDITCARDNUMBER": 46,
+    "B-CURRENCY": 54,
+    "B-CURRENCYCODE": 80,
+    "B-CURRENCYNAME": 69,
+    "B-CURRENCYSYMBOL": 71,
+    "B-DATE": 18,
+    "B-DISPLAYNAME": 112,
+    "B-EMAIL": 14,
+    "B-ETHEREUMADDRESS": 24,
+    "B-FIRSTNAME": 2,
+    "B-FULLNAME": 102,
+    "B-GENDER": 106,
+    "B-IBAN": 28,
+    "B-IP": 91,
+    "B-IPV4": 34,
+    "B-IPV6": 44,
+    "B-JOBAREA": 13,
+    "B-JOBDESCRIPTOR": 8,
+    "B-JOBTITLE": 9,
+    "B-JOBTYPE": 58,
+    "B-LASTNAME": 5,
+    "B-LITECOINADDRESS": 81,
+    "B-MAC": 83,
+    "B-MASKEDNUMBER": 72,
+    "B-MIDDLENAME": 4,
+    "B-NAME": 104,
+    "B-NEARBYGPSCOORDINATE": 110,
+    "B-NUMBER": 107,
+    "B-ORDINALDIRECTION": 115,
+    "B-PASSWORD": 56,
+    "B-PHONEIMEI": 64,
+    "B-PHONE_NUMBER": 74,
+    "B-PIN": 96,
+    "B-PREFIX": 0,
+    "B-SECONDARYADDRESS": 76,
+    "B-SEX": 93,
+    "B-SEXTYPE": 114,
+    "B-SSN": 78,
+    "B-STATE": 59,
+    "B-STREET": 94,
+    "B-STREETADDRESS": 36,
+    "B-SUFFIX": 50,
+    "B-TIME": 16,
+    "B-URL": 20,
+    "B-USERAGENT": 89,
+    "B-USERNAME": 42,
+    "B-VEHICLEVIN": 48,
+    "B-VEHICLEVRM": 62,
+    "B-ZIPCODE": 40,
+    "I-ACCOUNTNAME": 27,
+    "I-ACCOUNTNUMBER": 31,
+    "I-AMOUNT": 53,
+    "I-BIC": 33,
+    "I-BITCOINADDRESS": 23,
+    "I-BUILDINGNUMBER": 61,
+    "I-CITY": 39,
+    "I-COMPANY_NAME": 12,
+    "I-CREDITCARDCVV": 88,
+    "I-CREDITCARDISSUER": 86,
+    "I-CREDITCARDNUMBER": 47,
+    "I-CURRENCY": 55,
+    "I-CURRENCYCODE": 100,
+    "I-CURRENCYNAME": 70,
+    "I-CURRENCYSYMBOL": 101,
+    "I-DATE": 19,
+    "I-DISPLAYNAME": 113,
+    "I-EMAIL": 15,
+    "I-ETHEREUMADDRESS": 25,
+    "I-FIRSTNAME": 3,
+    "I-FULLNAME": 103,
+    "I-GENDER": 109,
+    "I-IBAN": 29,
+    "I-IP": 92,
+    "I-IPV4": 35,
+    "I-IPV6": 45,
+    "I-JOBAREA": 66,
+    "I-JOBTITLE": 10,
+    "I-JOBTYPE": 98,
+    "I-LASTNAME": 6,
+    "I-LITECOINADDRESS": 82,
+    "I-MAC": 84,
+    "I-MASKEDNUMBER": 73,
+    "I-MIDDLENAME": 99,
+    "I-NAME": 105,
+    "I-NEARBYGPSCOORDINATE": 111,
+    "I-NUMBER": 108,
+    "I-PASSWORD": 57,
+    "I-PHONEIMEI": 65,
+    "I-PHONE_NUMBER": 75,
+    "I-PIN": 97,
+    "I-PREFIX": 1,
+    "I-SECONDARYADDRESS": 77,
+    "I-SSN": 79,
+    "I-STATE": 67,
+    "I-STREET": 95,
+    "I-STREETADDRESS": 37,
+    "I-SUFFIX": 51,
+    "I-TIME": 17,
+    "I-URL": 21,
+    "I-USERAGENT": 90,
+    "I-USERNAME": 43,
+    "I-VEHICLEVIN": 49,
+    "I-VEHICLEVRM": 63,
+    "I-ZIPCODE": 41,
+    "O": 7
+  },
+  "max_position_embeddings": 512,
+  "model_type": "distilbert",
+  "n_heads": 12,
+  "n_layers": 6,
+  "pad_token_id": 0,
+  "qa_dropout": 0.1,
+  "seq_classif_dropout": 0.2,
+  "sinusoidal_pos_embds": false,
+  "tie_weights_": true,
+  "torch_dtype": "float32",
+  "transformers_version": "4.31.0",
+  "vocab_size": 30522
+}

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d693e8ad7a3acf6d30f2029728bec6a7fad75e6c93204217b840e8bb86619a5f
+size 265843173

special_tokens_map.json

@@ -0,0 +1,7 @@
+{
+  "cls_token": "[CLS]",
+  "mask_token": "[MASK]",
+  "pad_token": "[PAD]",
+  "sep_token": "[SEP]",
+  "unk_token": "[UNK]"
+}

tokenizer_config.json

@@ -0,0 +1,13 @@
+{
+  "clean_up_tokenization_spaces": true,
+  "cls_token": "[CLS]",
+  "do_lower_case": true,
+  "mask_token": "[MASK]",
+  "model_max_length": 512,
+  "pad_token": "[PAD]",
+  "sep_token": "[SEP]",
+  "strip_accents": null,
+  "tokenize_chinese_chars": true,
+  "tokenizer_class": "DistilBertTokenizer",
+  "unk_token": "[UNK]"
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c78f5eefe873f13228c3a9de122ce380b1553b0728c7c5301c6770dc9bbbf282
+size 4027