Upload folder using huggingface_hub
Browse files- gretel-pii-ready/span-deberta-v3-base/added_tokens.json +3 -0
- gretel-pii-ready/span-deberta-v3-base/checkpoint-25000/added_tokens.json +3 -0
- gretel-pii-ready/span-deberta-v3-base/checkpoint-25000/config.json +125 -0
- gretel-pii-ready/span-deberta-v3-base/checkpoint-25000/model.safetensors +3 -0
- gretel-pii-ready/span-deberta-v3-base/checkpoint-25000/optimizer.pt +3 -0
- gretel-pii-ready/span-deberta-v3-base/checkpoint-25000/rng_state.pth +3 -0
- gretel-pii-ready/span-deberta-v3-base/checkpoint-25000/scheduler.pt +3 -0
- gretel-pii-ready/span-deberta-v3-base/checkpoint-25000/special_tokens_map.json +15 -0
- gretel-pii-ready/span-deberta-v3-base/checkpoint-25000/spm.model +3 -0
- gretel-pii-ready/span-deberta-v3-base/checkpoint-25000/tokenizer.json +0 -0
- gretel-pii-ready/span-deberta-v3-base/checkpoint-25000/tokenizer_config.json +60 -0
- gretel-pii-ready/span-deberta-v3-base/checkpoint-25000/trainer_state.json +384 -0
- gretel-pii-ready/span-deberta-v3-base/checkpoint-25000/training_args.bin +3 -0
- gretel-pii-ready/span-deberta-v3-base/config.json +125 -0
- gretel-pii-ready/span-deberta-v3-base/pytorch_model.bin +3 -0
- gretel-pii-ready/span-deberta-v3-base/special_tokens_map.json +15 -0
- gretel-pii-ready/span-deberta-v3-base/spm.model +3 -0
- gretel-pii-ready/span-deberta-v3-base/tokenizer.json +0 -0
- gretel-pii-ready/span-deberta-v3-base/tokenizer_config.json +60 -0
- gretel-pii-ready/span-deberta-v3-base/training_meta.json +53 -0
gretel-pii-ready/span-deberta-v3-base/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"[MASK]": 128000
|
| 3 |
+
}
|
gretel-pii-ready/span-deberta-v3-base/checkpoint-25000/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"[MASK]": 128000
|
| 3 |
+
}
|
gretel-pii-ready/span-deberta-v3-base/checkpoint-25000/config.json
ADDED
|
@@ -0,0 +1,125 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"DebertaAdvancedSpanClassifier"
|
| 4 |
+
],
|
| 5 |
+
"attention_probs_dropout_prob": 0.1,
|
| 6 |
+
"dtype": "float32",
|
| 7 |
+
"hidden_act": "gelu",
|
| 8 |
+
"hidden_dropout_prob": 0.1,
|
| 9 |
+
"hidden_size": 768,
|
| 10 |
+
"id2label": {
|
| 11 |
+
"0": "O",
|
| 12 |
+
"1": "ACCOUNT_NUMBER",
|
| 13 |
+
"2": "ADDRESS",
|
| 14 |
+
"3": "API_KEY",
|
| 15 |
+
"4": "BANK_ROUTING_NUMBER",
|
| 16 |
+
"5": "BIOMETRIC_IDENTIFIER",
|
| 17 |
+
"6": "CERTIFICATE_LICENSE_NUMBER",
|
| 18 |
+
"7": "CITY",
|
| 19 |
+
"8": "COMPANY_NAME",
|
| 20 |
+
"9": "COORDINATE",
|
| 21 |
+
"10": "COUNTRY",
|
| 22 |
+
"11": "CREDIT_CARD_NUMBER",
|
| 23 |
+
"12": "CUSTOMER_ID",
|
| 24 |
+
"13": "CVV",
|
| 25 |
+
"14": "DATE",
|
| 26 |
+
"15": "DATE_OF_BIRTH",
|
| 27 |
+
"16": "DATE_TIME",
|
| 28 |
+
"17": "DEVICE_IDENTIFIER",
|
| 29 |
+
"18": "EMAIL",
|
| 30 |
+
"19": "EMPLOYEE_ID",
|
| 31 |
+
"20": "FIRST_NAME",
|
| 32 |
+
"21": "HEALTH_PLAN_BENEFICIARY_NUMBER",
|
| 33 |
+
"22": "IPV4",
|
| 34 |
+
"23": "IPV6",
|
| 35 |
+
"24": "LAST_NAME",
|
| 36 |
+
"25": "LICENSE_PLATE",
|
| 37 |
+
"26": "MEDICAL_RECORD_NUMBER",
|
| 38 |
+
"27": "NAME",
|
| 39 |
+
"28": "NATIONAL_ID",
|
| 40 |
+
"29": "PASSWORD",
|
| 41 |
+
"30": "PHONE_NUMBER",
|
| 42 |
+
"31": "PIN",
|
| 43 |
+
"32": "POSTCODE",
|
| 44 |
+
"33": "SSN",
|
| 45 |
+
"34": "STATE",
|
| 46 |
+
"35": "STREET_ADDRESS",
|
| 47 |
+
"36": "SWIFT_BIC",
|
| 48 |
+
"37": "TAX_ID",
|
| 49 |
+
"38": "TIME",
|
| 50 |
+
"39": "UNIQUE_IDENTIFIER",
|
| 51 |
+
"40": "URL",
|
| 52 |
+
"41": "USER_NAME",
|
| 53 |
+
"42": "VEHICLE_IDENTIFIER"
|
| 54 |
+
},
|
| 55 |
+
"initializer_range": 0.02,
|
| 56 |
+
"intermediate_size": 3072,
|
| 57 |
+
"label2id": {
|
| 58 |
+
"ACCOUNT_NUMBER": 1,
|
| 59 |
+
"ADDRESS": 2,
|
| 60 |
+
"API_KEY": 3,
|
| 61 |
+
"BANK_ROUTING_NUMBER": 4,
|
| 62 |
+
"BIOMETRIC_IDENTIFIER": 5,
|
| 63 |
+
"CERTIFICATE_LICENSE_NUMBER": 6,
|
| 64 |
+
"CITY": 7,
|
| 65 |
+
"COMPANY_NAME": 8,
|
| 66 |
+
"COORDINATE": 9,
|
| 67 |
+
"COUNTRY": 10,
|
| 68 |
+
"CREDIT_CARD_NUMBER": 11,
|
| 69 |
+
"CUSTOMER_ID": 12,
|
| 70 |
+
"CVV": 13,
|
| 71 |
+
"DATE": 14,
|
| 72 |
+
"DATE_OF_BIRTH": 15,
|
| 73 |
+
"DATE_TIME": 16,
|
| 74 |
+
"DEVICE_IDENTIFIER": 17,
|
| 75 |
+
"EMAIL": 18,
|
| 76 |
+
"EMPLOYEE_ID": 19,
|
| 77 |
+
"FIRST_NAME": 20,
|
| 78 |
+
"HEALTH_PLAN_BENEFICIARY_NUMBER": 21,
|
| 79 |
+
"IPV4": 22,
|
| 80 |
+
"IPV6": 23,
|
| 81 |
+
"LAST_NAME": 24,
|
| 82 |
+
"LICENSE_PLATE": 25,
|
| 83 |
+
"MEDICAL_RECORD_NUMBER": 26,
|
| 84 |
+
"NAME": 27,
|
| 85 |
+
"NATIONAL_ID": 28,
|
| 86 |
+
"O": 0,
|
| 87 |
+
"PASSWORD": 29,
|
| 88 |
+
"PHONE_NUMBER": 30,
|
| 89 |
+
"PIN": 31,
|
| 90 |
+
"POSTCODE": 32,
|
| 91 |
+
"SSN": 33,
|
| 92 |
+
"STATE": 34,
|
| 93 |
+
"STREET_ADDRESS": 35,
|
| 94 |
+
"SWIFT_BIC": 36,
|
| 95 |
+
"TAX_ID": 37,
|
| 96 |
+
"TIME": 38,
|
| 97 |
+
"UNIQUE_IDENTIFIER": 39,
|
| 98 |
+
"URL": 40,
|
| 99 |
+
"USER_NAME": 41,
|
| 100 |
+
"VEHICLE_IDENTIFIER": 42
|
| 101 |
+
},
|
| 102 |
+
"layer_norm_eps": 1e-07,
|
| 103 |
+
"legacy": true,
|
| 104 |
+
"max_position_embeddings": 512,
|
| 105 |
+
"max_relative_positions": -1,
|
| 106 |
+
"model_type": "deberta-v2",
|
| 107 |
+
"norm_rel_ebd": "layer_norm",
|
| 108 |
+
"num_attention_heads": 12,
|
| 109 |
+
"num_hidden_layers": 12,
|
| 110 |
+
"pad_token_id": 0,
|
| 111 |
+
"pooler_dropout": 0,
|
| 112 |
+
"pooler_hidden_act": "gelu",
|
| 113 |
+
"pooler_hidden_size": 768,
|
| 114 |
+
"pos_att_type": [
|
| 115 |
+
"p2c",
|
| 116 |
+
"c2p"
|
| 117 |
+
],
|
| 118 |
+
"position_biased_input": false,
|
| 119 |
+
"position_buckets": 256,
|
| 120 |
+
"relative_attention": true,
|
| 121 |
+
"share_att_key": true,
|
| 122 |
+
"transformers_version": "4.57.6",
|
| 123 |
+
"type_vocab_size": 0,
|
| 124 |
+
"vocab_size": 128100
|
| 125 |
+
}
|
gretel-pii-ready/span-deberta-v3-base/checkpoint-25000/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e5507d52e14019c9bb67154d941f3d94536641628d87ceaef2459c6b08d2815e
|
| 3 |
+
size 741430532
|
gretel-pii-ready/span-deberta-v3-base/checkpoint-25000/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8ea0a071e9b787109a9a56e0275ba209c10c5ef0410933ad4944ebfe2419cd65
|
| 3 |
+
size 1482984715
|
gretel-pii-ready/span-deberta-v3-base/checkpoint-25000/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:beee111b74fd02e9f4d9e6c628d384455298e0e9bc6b36ff629a694d6a14e268
|
| 3 |
+
size 14645
|
gretel-pii-ready/span-deberta-v3-base/checkpoint-25000/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2d03520265763da43353400b101c59aafa3767ece20ed1e902dec77e4c0e03d6
|
| 3 |
+
size 1465
|
gretel-pii-ready/span-deberta-v3-base/checkpoint-25000/special_tokens_map.json
ADDED
|
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "[CLS]",
|
| 3 |
+
"cls_token": "[CLS]",
|
| 4 |
+
"eos_token": "[SEP]",
|
| 5 |
+
"mask_token": "[MASK]",
|
| 6 |
+
"pad_token": "[PAD]",
|
| 7 |
+
"sep_token": "[SEP]",
|
| 8 |
+
"unk_token": {
|
| 9 |
+
"content": "[UNK]",
|
| 10 |
+
"lstrip": false,
|
| 11 |
+
"normalized": true,
|
| 12 |
+
"rstrip": false,
|
| 13 |
+
"single_word": false
|
| 14 |
+
}
|
| 15 |
+
}
|
gretel-pii-ready/span-deberta-v3-base/checkpoint-25000/spm.model
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c679fbf93643d19aab7ee10c0b99e460bdbc02fedf34b92b05af343b4af586fd
|
| 3 |
+
size 2464616
|
gretel-pii-ready/span-deberta-v3-base/checkpoint-25000/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
gretel-pii-ready/span-deberta-v3-base/checkpoint-25000/tokenizer_config.json
ADDED
|
@@ -0,0 +1,60 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": true,
|
| 3 |
+
"added_tokens_decoder": {
|
| 4 |
+
"0": {
|
| 5 |
+
"content": "[PAD]",
|
| 6 |
+
"lstrip": false,
|
| 7 |
+
"normalized": false,
|
| 8 |
+
"rstrip": false,
|
| 9 |
+
"single_word": false,
|
| 10 |
+
"special": true
|
| 11 |
+
},
|
| 12 |
+
"1": {
|
| 13 |
+
"content": "[CLS]",
|
| 14 |
+
"lstrip": false,
|
| 15 |
+
"normalized": false,
|
| 16 |
+
"rstrip": false,
|
| 17 |
+
"single_word": false,
|
| 18 |
+
"special": true
|
| 19 |
+
},
|
| 20 |
+
"2": {
|
| 21 |
+
"content": "[SEP]",
|
| 22 |
+
"lstrip": false,
|
| 23 |
+
"normalized": false,
|
| 24 |
+
"rstrip": false,
|
| 25 |
+
"single_word": false,
|
| 26 |
+
"special": true
|
| 27 |
+
},
|
| 28 |
+
"3": {
|
| 29 |
+
"content": "[UNK]",
|
| 30 |
+
"lstrip": false,
|
| 31 |
+
"normalized": true,
|
| 32 |
+
"rstrip": false,
|
| 33 |
+
"single_word": false,
|
| 34 |
+
"special": true
|
| 35 |
+
},
|
| 36 |
+
"128000": {
|
| 37 |
+
"content": "[MASK]",
|
| 38 |
+
"lstrip": false,
|
| 39 |
+
"normalized": false,
|
| 40 |
+
"rstrip": false,
|
| 41 |
+
"single_word": false,
|
| 42 |
+
"special": true
|
| 43 |
+
}
|
| 44 |
+
},
|
| 45 |
+
"bos_token": "[CLS]",
|
| 46 |
+
"clean_up_tokenization_spaces": false,
|
| 47 |
+
"cls_token": "[CLS]",
|
| 48 |
+
"do_lower_case": false,
|
| 49 |
+
"eos_token": "[SEP]",
|
| 50 |
+
"extra_special_tokens": {},
|
| 51 |
+
"mask_token": "[MASK]",
|
| 52 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 53 |
+
"pad_token": "[PAD]",
|
| 54 |
+
"sep_token": "[SEP]",
|
| 55 |
+
"sp_model_kwargs": {},
|
| 56 |
+
"split_by_punct": false,
|
| 57 |
+
"tokenizer_class": "DebertaV2Tokenizer",
|
| 58 |
+
"unk_token": "[UNK]",
|
| 59 |
+
"vocab_type": "spm"
|
| 60 |
+
}
|
gretel-pii-ready/span-deberta-v3-base/checkpoint-25000/trainer_state.json
ADDED
|
@@ -0,0 +1,384 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 1.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 25000,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.02,
|
| 14 |
+
"grad_norm": 0.1428048014640808,
|
| 15 |
+
"learning_rate": 9.98e-07,
|
| 16 |
+
"loss": 0.0632,
|
| 17 |
+
"step": 500
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.04,
|
| 21 |
+
"grad_norm": 0.033088039606809616,
|
| 22 |
+
"learning_rate": 1.998e-06,
|
| 23 |
+
"loss": 0.0215,
|
| 24 |
+
"step": 1000
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.06,
|
| 28 |
+
"grad_norm": 0.04278671368956566,
|
| 29 |
+
"learning_rate": 2.9980000000000003e-06,
|
| 30 |
+
"loss": 0.0141,
|
| 31 |
+
"step": 1500
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.08,
|
| 35 |
+
"grad_norm": 0.043109361082315445,
|
| 36 |
+
"learning_rate": 3.9980000000000005e-06,
|
| 37 |
+
"loss": 0.0097,
|
| 38 |
+
"step": 2000
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.1,
|
| 42 |
+
"grad_norm": 0.02095414139330387,
|
| 43 |
+
"learning_rate": 4.998e-06,
|
| 44 |
+
"loss": 0.0066,
|
| 45 |
+
"step": 2500
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.12,
|
| 49 |
+
"grad_norm": 0.02322233095765114,
|
| 50 |
+
"learning_rate": 4.889111111111112e-06,
|
| 51 |
+
"loss": 0.0046,
|
| 52 |
+
"step": 3000
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.14,
|
| 56 |
+
"grad_norm": 0.023266147822141647,
|
| 57 |
+
"learning_rate": 4.778000000000001e-06,
|
| 58 |
+
"loss": 0.0038,
|
| 59 |
+
"step": 3500
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 0.16,
|
| 63 |
+
"grad_norm": 0.01793760620057583,
|
| 64 |
+
"learning_rate": 4.666888888888889e-06,
|
| 65 |
+
"loss": 0.0028,
|
| 66 |
+
"step": 4000
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 0.18,
|
| 70 |
+
"grad_norm": 0.01407337561249733,
|
| 71 |
+
"learning_rate": 4.5557777777777785e-06,
|
| 72 |
+
"loss": 0.0021,
|
| 73 |
+
"step": 4500
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 0.2,
|
| 77 |
+
"grad_norm": 0.048996467143297195,
|
| 78 |
+
"learning_rate": 4.444666666666667e-06,
|
| 79 |
+
"loss": 0.0018,
|
| 80 |
+
"step": 5000
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.22,
|
| 84 |
+
"grad_norm": 0.016084177419543266,
|
| 85 |
+
"learning_rate": 4.3335555555555555e-06,
|
| 86 |
+
"loss": 0.0018,
|
| 87 |
+
"step": 5500
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.24,
|
| 91 |
+
"grad_norm": 0.017788106575608253,
|
| 92 |
+
"learning_rate": 4.222444444444444e-06,
|
| 93 |
+
"loss": 0.0016,
|
| 94 |
+
"step": 6000
|
| 95 |
+
},
|
| 96 |
+
{
|
| 97 |
+
"epoch": 0.26,
|
| 98 |
+
"grad_norm": 0.03554818406701088,
|
| 99 |
+
"learning_rate": 4.111333333333334e-06,
|
| 100 |
+
"loss": 0.0014,
|
| 101 |
+
"step": 6500
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 0.28,
|
| 105 |
+
"grad_norm": 0.010203654877841473,
|
| 106 |
+
"learning_rate": 4.000222222222222e-06,
|
| 107 |
+
"loss": 0.0014,
|
| 108 |
+
"step": 7000
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"epoch": 0.3,
|
| 112 |
+
"grad_norm": 0.019548939540982246,
|
| 113 |
+
"learning_rate": 3.889111111111111e-06,
|
| 114 |
+
"loss": 0.0012,
|
| 115 |
+
"step": 7500
|
| 116 |
+
},
|
| 117 |
+
{
|
| 118 |
+
"epoch": 0.32,
|
| 119 |
+
"grad_norm": 0.067998968064785,
|
| 120 |
+
"learning_rate": 3.7780000000000005e-06,
|
| 121 |
+
"loss": 0.001,
|
| 122 |
+
"step": 8000
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.34,
|
| 126 |
+
"grad_norm": 0.04204774275422096,
|
| 127 |
+
"learning_rate": 3.6668888888888894e-06,
|
| 128 |
+
"loss": 0.001,
|
| 129 |
+
"step": 8500
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.36,
|
| 133 |
+
"grad_norm": 0.0033883412834256887,
|
| 134 |
+
"learning_rate": 3.555777777777778e-06,
|
| 135 |
+
"loss": 0.001,
|
| 136 |
+
"step": 9000
|
| 137 |
+
},
|
| 138 |
+
{
|
| 139 |
+
"epoch": 0.38,
|
| 140 |
+
"grad_norm": 0.04815370589494705,
|
| 141 |
+
"learning_rate": 3.444666666666667e-06,
|
| 142 |
+
"loss": 0.001,
|
| 143 |
+
"step": 9500
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"epoch": 0.4,
|
| 147 |
+
"grad_norm": 0.009633071720600128,
|
| 148 |
+
"learning_rate": 3.333555555555556e-06,
|
| 149 |
+
"loss": 0.0008,
|
| 150 |
+
"step": 10000
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"epoch": 0.42,
|
| 154 |
+
"grad_norm": 0.010944943875074387,
|
| 155 |
+
"learning_rate": 3.2224444444444447e-06,
|
| 156 |
+
"loss": 0.0008,
|
| 157 |
+
"step": 10500
|
| 158 |
+
},
|
| 159 |
+
{
|
| 160 |
+
"epoch": 0.44,
|
| 161 |
+
"grad_norm": 0.037118714302778244,
|
| 162 |
+
"learning_rate": 3.1113333333333336e-06,
|
| 163 |
+
"loss": 0.0008,
|
| 164 |
+
"step": 11000
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.46,
|
| 168 |
+
"grad_norm": 0.017736153677105904,
|
| 169 |
+
"learning_rate": 3.0002222222222225e-06,
|
| 170 |
+
"loss": 0.0008,
|
| 171 |
+
"step": 11500
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.48,
|
| 175 |
+
"grad_norm": 0.012890961021184921,
|
| 176 |
+
"learning_rate": 2.8891111111111114e-06,
|
| 177 |
+
"loss": 0.0008,
|
| 178 |
+
"step": 12000
|
| 179 |
+
},
|
| 180 |
+
{
|
| 181 |
+
"epoch": 0.5,
|
| 182 |
+
"grad_norm": 0.012737075798213482,
|
| 183 |
+
"learning_rate": 2.7780000000000003e-06,
|
| 184 |
+
"loss": 0.0008,
|
| 185 |
+
"step": 12500
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 0.52,
|
| 189 |
+
"grad_norm": 0.027506740763783455,
|
| 190 |
+
"learning_rate": 2.666888888888889e-06,
|
| 191 |
+
"loss": 0.0007,
|
| 192 |
+
"step": 13000
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 0.54,
|
| 196 |
+
"grad_norm": 0.010766581632196903,
|
| 197 |
+
"learning_rate": 2.555777777777778e-06,
|
| 198 |
+
"loss": 0.0006,
|
| 199 |
+
"step": 13500
|
| 200 |
+
},
|
| 201 |
+
{
|
| 202 |
+
"epoch": 0.56,
|
| 203 |
+
"grad_norm": 0.011902324855327606,
|
| 204 |
+
"learning_rate": 2.4446666666666667e-06,
|
| 205 |
+
"loss": 0.0007,
|
| 206 |
+
"step": 14000
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.58,
|
| 210 |
+
"grad_norm": 0.002812189282849431,
|
| 211 |
+
"learning_rate": 2.333555555555556e-06,
|
| 212 |
+
"loss": 0.0006,
|
| 213 |
+
"step": 14500
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 0.6,
|
| 217 |
+
"grad_norm": 0.005815847776830196,
|
| 218 |
+
"learning_rate": 2.2224444444444445e-06,
|
| 219 |
+
"loss": 0.0007,
|
| 220 |
+
"step": 15000
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"epoch": 0.62,
|
| 224 |
+
"grad_norm": 0.01781208999454975,
|
| 225 |
+
"learning_rate": 2.1113333333333334e-06,
|
| 226 |
+
"loss": 0.0007,
|
| 227 |
+
"step": 15500
|
| 228 |
+
},
|
| 229 |
+
{
|
| 230 |
+
"epoch": 0.64,
|
| 231 |
+
"grad_norm": 0.01837192103266716,
|
| 232 |
+
"learning_rate": 2.0002222222222223e-06,
|
| 233 |
+
"loss": 0.0006,
|
| 234 |
+
"step": 16000
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 0.66,
|
| 238 |
+
"grad_norm": 0.04189838469028473,
|
| 239 |
+
"learning_rate": 1.8891111111111113e-06,
|
| 240 |
+
"loss": 0.0006,
|
| 241 |
+
"step": 16500
|
| 242 |
+
},
|
| 243 |
+
{
|
| 244 |
+
"epoch": 0.68,
|
| 245 |
+
"grad_norm": 0.0050574480555951595,
|
| 246 |
+
"learning_rate": 1.7780000000000004e-06,
|
| 247 |
+
"loss": 0.0006,
|
| 248 |
+
"step": 17000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.7,
|
| 252 |
+
"grad_norm": 0.017372770234942436,
|
| 253 |
+
"learning_rate": 1.666888888888889e-06,
|
| 254 |
+
"loss": 0.0006,
|
| 255 |
+
"step": 17500
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 0.72,
|
| 259 |
+
"grad_norm": 0.006160199176520109,
|
| 260 |
+
"learning_rate": 1.555777777777778e-06,
|
| 261 |
+
"loss": 0.0005,
|
| 262 |
+
"step": 18000
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"epoch": 0.74,
|
| 266 |
+
"grad_norm": 0.002297651022672653,
|
| 267 |
+
"learning_rate": 1.4446666666666667e-06,
|
| 268 |
+
"loss": 0.0005,
|
| 269 |
+
"step": 18500
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 0.76,
|
| 273 |
+
"grad_norm": 0.010377667844295502,
|
| 274 |
+
"learning_rate": 1.3335555555555556e-06,
|
| 275 |
+
"loss": 0.0006,
|
| 276 |
+
"step": 19000
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 0.78,
|
| 280 |
+
"grad_norm": 0.007855025120079517,
|
| 281 |
+
"learning_rate": 1.2224444444444445e-06,
|
| 282 |
+
"loss": 0.0005,
|
| 283 |
+
"step": 19500
|
| 284 |
+
},
|
| 285 |
+
{
|
| 286 |
+
"epoch": 0.8,
|
| 287 |
+
"grad_norm": 0.0111708277836442,
|
| 288 |
+
"learning_rate": 1.1113333333333335e-06,
|
| 289 |
+
"loss": 0.0005,
|
| 290 |
+
"step": 20000
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.82,
|
| 294 |
+
"grad_norm": 0.0022401015739887953,
|
| 295 |
+
"learning_rate": 1.0002222222222224e-06,
|
| 296 |
+
"loss": 0.0005,
|
| 297 |
+
"step": 20500
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.84,
|
| 301 |
+
"grad_norm": 0.01948193833231926,
|
| 302 |
+
"learning_rate": 8.891111111111112e-07,
|
| 303 |
+
"loss": 0.0006,
|
| 304 |
+
"step": 21000
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 0.86,
|
| 308 |
+
"grad_norm": 0.04793281480669975,
|
| 309 |
+
"learning_rate": 7.78e-07,
|
| 310 |
+
"loss": 0.0006,
|
| 311 |
+
"step": 21500
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 0.88,
|
| 315 |
+
"grad_norm": 0.015651477500796318,
|
| 316 |
+
"learning_rate": 6.66888888888889e-07,
|
| 317 |
+
"loss": 0.0005,
|
| 318 |
+
"step": 22000
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"epoch": 0.9,
|
| 322 |
+
"grad_norm": 0.04631660506129265,
|
| 323 |
+
"learning_rate": 5.557777777777778e-07,
|
| 324 |
+
"loss": 0.0005,
|
| 325 |
+
"step": 22500
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 0.92,
|
| 329 |
+
"grad_norm": 0.0008624398615211248,
|
| 330 |
+
"learning_rate": 4.446666666666667e-07,
|
| 331 |
+
"loss": 0.0005,
|
| 332 |
+
"step": 23000
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.94,
|
| 336 |
+
"grad_norm": 0.001822498976252973,
|
| 337 |
+
"learning_rate": 3.3355555555555557e-07,
|
| 338 |
+
"loss": 0.0006,
|
| 339 |
+
"step": 23500
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 0.96,
|
| 343 |
+
"grad_norm": 0.003358706133440137,
|
| 344 |
+
"learning_rate": 2.2244444444444446e-07,
|
| 345 |
+
"loss": 0.0005,
|
| 346 |
+
"step": 24000
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 0.98,
|
| 350 |
+
"grad_norm": 0.006471550557762384,
|
| 351 |
+
"learning_rate": 1.1133333333333335e-07,
|
| 352 |
+
"loss": 0.0005,
|
| 353 |
+
"step": 24500
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 1.0,
|
| 357 |
+
"grad_norm": 0.01957268826663494,
|
| 358 |
+
"learning_rate": 2.2222222222222224e-10,
|
| 359 |
+
"loss": 0.0005,
|
| 360 |
+
"step": 25000
|
| 361 |
+
}
|
| 362 |
+
],
|
| 363 |
+
"logging_steps": 500,
|
| 364 |
+
"max_steps": 25000,
|
| 365 |
+
"num_input_tokens_seen": 0,
|
| 366 |
+
"num_train_epochs": 1,
|
| 367 |
+
"save_steps": 500,
|
| 368 |
+
"stateful_callbacks": {
|
| 369 |
+
"TrainerControl": {
|
| 370 |
+
"args": {
|
| 371 |
+
"should_epoch_stop": false,
|
| 372 |
+
"should_evaluate": false,
|
| 373 |
+
"should_log": false,
|
| 374 |
+
"should_save": true,
|
| 375 |
+
"should_training_stop": true
|
| 376 |
+
},
|
| 377 |
+
"attributes": {}
|
| 378 |
+
}
|
| 379 |
+
},
|
| 380 |
+
"total_flos": 2262335509104888.0,
|
| 381 |
+
"train_batch_size": 2,
|
| 382 |
+
"trial_name": null,
|
| 383 |
+
"trial_params": null
|
| 384 |
+
}
|
gretel-pii-ready/span-deberta-v3-base/checkpoint-25000/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:210ae1f29f955741f287e19238cc0aef815c07ce59da5cd3f4a01f1f8f1e16a7
|
| 3 |
+
size 5841
|
gretel-pii-ready/span-deberta-v3-base/config.json
ADDED
|
@@ -0,0 +1,125 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"DebertaAdvancedSpanClassifier"
|
| 4 |
+
],
|
| 5 |
+
"attention_probs_dropout_prob": 0.1,
|
| 6 |
+
"dtype": "float32",
|
| 7 |
+
"hidden_act": "gelu",
|
| 8 |
+
"hidden_dropout_prob": 0.1,
|
| 9 |
+
"hidden_size": 768,
|
| 10 |
+
"id2label": {
|
| 11 |
+
"0": "O",
|
| 12 |
+
"1": "ACCOUNT_NUMBER",
|
| 13 |
+
"2": "ADDRESS",
|
| 14 |
+
"3": "API_KEY",
|
| 15 |
+
"4": "BANK_ROUTING_NUMBER",
|
| 16 |
+
"5": "BIOMETRIC_IDENTIFIER",
|
| 17 |
+
"6": "CERTIFICATE_LICENSE_NUMBER",
|
| 18 |
+
"7": "CITY",
|
| 19 |
+
"8": "COMPANY_NAME",
|
| 20 |
+
"9": "COORDINATE",
|
| 21 |
+
"10": "COUNTRY",
|
| 22 |
+
"11": "CREDIT_CARD_NUMBER",
|
| 23 |
+
"12": "CUSTOMER_ID",
|
| 24 |
+
"13": "CVV",
|
| 25 |
+
"14": "DATE",
|
| 26 |
+
"15": "DATE_OF_BIRTH",
|
| 27 |
+
"16": "DATE_TIME",
|
| 28 |
+
"17": "DEVICE_IDENTIFIER",
|
| 29 |
+
"18": "EMAIL",
|
| 30 |
+
"19": "EMPLOYEE_ID",
|
| 31 |
+
"20": "FIRST_NAME",
|
| 32 |
+
"21": "HEALTH_PLAN_BENEFICIARY_NUMBER",
|
| 33 |
+
"22": "IPV4",
|
| 34 |
+
"23": "IPV6",
|
| 35 |
+
"24": "LAST_NAME",
|
| 36 |
+
"25": "LICENSE_PLATE",
|
| 37 |
+
"26": "MEDICAL_RECORD_NUMBER",
|
| 38 |
+
"27": "NAME",
|
| 39 |
+
"28": "NATIONAL_ID",
|
| 40 |
+
"29": "PASSWORD",
|
| 41 |
+
"30": "PHONE_NUMBER",
|
| 42 |
+
"31": "PIN",
|
| 43 |
+
"32": "POSTCODE",
|
| 44 |
+
"33": "SSN",
|
| 45 |
+
"34": "STATE",
|
| 46 |
+
"35": "STREET_ADDRESS",
|
| 47 |
+
"36": "SWIFT_BIC",
|
| 48 |
+
"37": "TAX_ID",
|
| 49 |
+
"38": "TIME",
|
| 50 |
+
"39": "UNIQUE_IDENTIFIER",
|
| 51 |
+
"40": "URL",
|
| 52 |
+
"41": "USER_NAME",
|
| 53 |
+
"42": "VEHICLE_IDENTIFIER"
|
| 54 |
+
},
|
| 55 |
+
"initializer_range": 0.02,
|
| 56 |
+
"intermediate_size": 3072,
|
| 57 |
+
"label2id": {
|
| 58 |
+
"ACCOUNT_NUMBER": 1,
|
| 59 |
+
"ADDRESS": 2,
|
| 60 |
+
"API_KEY": 3,
|
| 61 |
+
"BANK_ROUTING_NUMBER": 4,
|
| 62 |
+
"BIOMETRIC_IDENTIFIER": 5,
|
| 63 |
+
"CERTIFICATE_LICENSE_NUMBER": 6,
|
| 64 |
+
"CITY": 7,
|
| 65 |
+
"COMPANY_NAME": 8,
|
| 66 |
+
"COORDINATE": 9,
|
| 67 |
+
"COUNTRY": 10,
|
| 68 |
+
"CREDIT_CARD_NUMBER": 11,
|
| 69 |
+
"CUSTOMER_ID": 12,
|
| 70 |
+
"CVV": 13,
|
| 71 |
+
"DATE": 14,
|
| 72 |
+
"DATE_OF_BIRTH": 15,
|
| 73 |
+
"DATE_TIME": 16,
|
| 74 |
+
"DEVICE_IDENTIFIER": 17,
|
| 75 |
+
"EMAIL": 18,
|
| 76 |
+
"EMPLOYEE_ID": 19,
|
| 77 |
+
"FIRST_NAME": 20,
|
| 78 |
+
"HEALTH_PLAN_BENEFICIARY_NUMBER": 21,
|
| 79 |
+
"IPV4": 22,
|
| 80 |
+
"IPV6": 23,
|
| 81 |
+
"LAST_NAME": 24,
|
| 82 |
+
"LICENSE_PLATE": 25,
|
| 83 |
+
"MEDICAL_RECORD_NUMBER": 26,
|
| 84 |
+
"NAME": 27,
|
| 85 |
+
"NATIONAL_ID": 28,
|
| 86 |
+
"O": 0,
|
| 87 |
+
"PASSWORD": 29,
|
| 88 |
+
"PHONE_NUMBER": 30,
|
| 89 |
+
"PIN": 31,
|
| 90 |
+
"POSTCODE": 32,
|
| 91 |
+
"SSN": 33,
|
| 92 |
+
"STATE": 34,
|
| 93 |
+
"STREET_ADDRESS": 35,
|
| 94 |
+
"SWIFT_BIC": 36,
|
| 95 |
+
"TAX_ID": 37,
|
| 96 |
+
"TIME": 38,
|
| 97 |
+
"UNIQUE_IDENTIFIER": 39,
|
| 98 |
+
"URL": 40,
|
| 99 |
+
"USER_NAME": 41,
|
| 100 |
+
"VEHICLE_IDENTIFIER": 42
|
| 101 |
+
},
|
| 102 |
+
"layer_norm_eps": 1e-07,
|
| 103 |
+
"legacy": true,
|
| 104 |
+
"max_position_embeddings": 512,
|
| 105 |
+
"max_relative_positions": -1,
|
| 106 |
+
"model_type": "deberta-v2",
|
| 107 |
+
"norm_rel_ebd": "layer_norm",
|
| 108 |
+
"num_attention_heads": 12,
|
| 109 |
+
"num_hidden_layers": 12,
|
| 110 |
+
"pad_token_id": 0,
|
| 111 |
+
"pooler_dropout": 0,
|
| 112 |
+
"pooler_hidden_act": "gelu",
|
| 113 |
+
"pooler_hidden_size": 768,
|
| 114 |
+
"pos_att_type": [
|
| 115 |
+
"p2c",
|
| 116 |
+
"c2p"
|
| 117 |
+
],
|
| 118 |
+
"position_biased_input": false,
|
| 119 |
+
"position_buckets": 256,
|
| 120 |
+
"relative_attention": true,
|
| 121 |
+
"share_att_key": true,
|
| 122 |
+
"transformers_version": "4.57.6",
|
| 123 |
+
"type_vocab_size": 0,
|
| 124 |
+
"vocab_size": 128100
|
| 125 |
+
}
|
gretel-pii-ready/span-deberta-v3-base/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a4e7f4e2df9afdf6a64fa175527b221f35496d699026f19a07430e814a85506f
|
| 3 |
+
size 741492431
|
gretel-pii-ready/span-deberta-v3-base/special_tokens_map.json
ADDED
|
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "[CLS]",
|
| 3 |
+
"cls_token": "[CLS]",
|
| 4 |
+
"eos_token": "[SEP]",
|
| 5 |
+
"mask_token": "[MASK]",
|
| 6 |
+
"pad_token": "[PAD]",
|
| 7 |
+
"sep_token": "[SEP]",
|
| 8 |
+
"unk_token": {
|
| 9 |
+
"content": "[UNK]",
|
| 10 |
+
"lstrip": false,
|
| 11 |
+
"normalized": true,
|
| 12 |
+
"rstrip": false,
|
| 13 |
+
"single_word": false
|
| 14 |
+
}
|
| 15 |
+
}
|
gretel-pii-ready/span-deberta-v3-base/spm.model
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c679fbf93643d19aab7ee10c0b99e460bdbc02fedf34b92b05af343b4af586fd
|
| 3 |
+
size 2464616
|
gretel-pii-ready/span-deberta-v3-base/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
gretel-pii-ready/span-deberta-v3-base/tokenizer_config.json
ADDED
|
@@ -0,0 +1,60 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": true,
|
| 3 |
+
"added_tokens_decoder": {
|
| 4 |
+
"0": {
|
| 5 |
+
"content": "[PAD]",
|
| 6 |
+
"lstrip": false,
|
| 7 |
+
"normalized": false,
|
| 8 |
+
"rstrip": false,
|
| 9 |
+
"single_word": false,
|
| 10 |
+
"special": true
|
| 11 |
+
},
|
| 12 |
+
"1": {
|
| 13 |
+
"content": "[CLS]",
|
| 14 |
+
"lstrip": false,
|
| 15 |
+
"normalized": false,
|
| 16 |
+
"rstrip": false,
|
| 17 |
+
"single_word": false,
|
| 18 |
+
"special": true
|
| 19 |
+
},
|
| 20 |
+
"2": {
|
| 21 |
+
"content": "[SEP]",
|
| 22 |
+
"lstrip": false,
|
| 23 |
+
"normalized": false,
|
| 24 |
+
"rstrip": false,
|
| 25 |
+
"single_word": false,
|
| 26 |
+
"special": true
|
| 27 |
+
},
|
| 28 |
+
"3": {
|
| 29 |
+
"content": "[UNK]",
|
| 30 |
+
"lstrip": false,
|
| 31 |
+
"normalized": true,
|
| 32 |
+
"rstrip": false,
|
| 33 |
+
"single_word": false,
|
| 34 |
+
"special": true
|
| 35 |
+
},
|
| 36 |
+
"128000": {
|
| 37 |
+
"content": "[MASK]",
|
| 38 |
+
"lstrip": false,
|
| 39 |
+
"normalized": false,
|
| 40 |
+
"rstrip": false,
|
| 41 |
+
"single_word": false,
|
| 42 |
+
"special": true
|
| 43 |
+
}
|
| 44 |
+
},
|
| 45 |
+
"bos_token": "[CLS]",
|
| 46 |
+
"clean_up_tokenization_spaces": false,
|
| 47 |
+
"cls_token": "[CLS]",
|
| 48 |
+
"do_lower_case": false,
|
| 49 |
+
"eos_token": "[SEP]",
|
| 50 |
+
"extra_special_tokens": {},
|
| 51 |
+
"mask_token": "[MASK]",
|
| 52 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 53 |
+
"pad_token": "[PAD]",
|
| 54 |
+
"sep_token": "[SEP]",
|
| 55 |
+
"sp_model_kwargs": {},
|
| 56 |
+
"split_by_punct": false,
|
| 57 |
+
"tokenizer_class": "DebertaV2Tokenizer",
|
| 58 |
+
"unk_token": "[UNK]",
|
| 59 |
+
"vocab_type": "spm"
|
| 60 |
+
}
|
gretel-pii-ready/span-deberta-v3-base/training_meta.json
ADDED
|
@@ -0,0 +1,53 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architecture": "span_based",
|
| 3 |
+
"base_model": "microsoft/deberta-v3-base",
|
| 4 |
+
"dataset": "gretel-pii-ready",
|
| 5 |
+
"label_list": [
|
| 6 |
+
"O",
|
| 7 |
+
"ACCOUNT_NUMBER",
|
| 8 |
+
"ADDRESS",
|
| 9 |
+
"API_KEY",
|
| 10 |
+
"BANK_ROUTING_NUMBER",
|
| 11 |
+
"BIOMETRIC_IDENTIFIER",
|
| 12 |
+
"CERTIFICATE_LICENSE_NUMBER",
|
| 13 |
+
"CITY",
|
| 14 |
+
"COMPANY_NAME",
|
| 15 |
+
"COORDINATE",
|
| 16 |
+
"COUNTRY",
|
| 17 |
+
"CREDIT_CARD_NUMBER",
|
| 18 |
+
"CUSTOMER_ID",
|
| 19 |
+
"CVV",
|
| 20 |
+
"DATE",
|
| 21 |
+
"DATE_OF_BIRTH",
|
| 22 |
+
"DATE_TIME",
|
| 23 |
+
"DEVICE_IDENTIFIER",
|
| 24 |
+
"EMAIL",
|
| 25 |
+
"EMPLOYEE_ID",
|
| 26 |
+
"FIRST_NAME",
|
| 27 |
+
"HEALTH_PLAN_BENEFICIARY_NUMBER",
|
| 28 |
+
"IPV4",
|
| 29 |
+
"IPV6",
|
| 30 |
+
"LAST_NAME",
|
| 31 |
+
"LICENSE_PLATE",
|
| 32 |
+
"MEDICAL_RECORD_NUMBER",
|
| 33 |
+
"NAME",
|
| 34 |
+
"NATIONAL_ID",
|
| 35 |
+
"PASSWORD",
|
| 36 |
+
"PHONE_NUMBER",
|
| 37 |
+
"PIN",
|
| 38 |
+
"POSTCODE",
|
| 39 |
+
"SSN",
|
| 40 |
+
"STATE",
|
| 41 |
+
"STREET_ADDRESS",
|
| 42 |
+
"SWIFT_BIC",
|
| 43 |
+
"TAX_ID",
|
| 44 |
+
"TIME",
|
| 45 |
+
"UNIQUE_IDENTIFIER",
|
| 46 |
+
"URL",
|
| 47 |
+
"USER_NAME",
|
| 48 |
+
"VEHICLE_IDENTIFIER"
|
| 49 |
+
],
|
| 50 |
+
"max_seq_length": 512,
|
| 51 |
+
"max_span_width": 30,
|
| 52 |
+
"timestamp": "2026-04-05T04:04:08.243730"
|
| 53 |
+
}
|