Hugging Face's logo

PuxAI
/
Internal-PII-Models

Safetensors
Model card Files
xet
 Community
PuxAI commited on
Commit
0ac7195
·
verified ·
1 Parent(s): f230a75

Upload folder using huggingface_hub

Browse files
Files changed (10) hide show
  1. Internal-Run-EN/SpanBased-DeBERTa/checkpoint-5788/config.json +157 -0
  2. Internal-Run-EN/SpanBased-DeBERTa/checkpoint-5788/model.safetensors +3 -0
  3. Internal-Run-EN/SpanBased-DeBERTa/checkpoint-5788/optimizer.pt +3 -0
  4. Internal-Run-EN/SpanBased-DeBERTa/checkpoint-5788/rng_state.pth +3 -0
  5. Internal-Run-EN/SpanBased-DeBERTa/checkpoint-5788/scheduler.pt +3 -0
  6. Internal-Run-EN/SpanBased-DeBERTa/checkpoint-5788/tokenizer.json +0 -0
  7. Internal-Run-EN/SpanBased-DeBERTa/checkpoint-5788/tokenizer_config.json +23 -0
  8. Internal-Run-EN/SpanBased-DeBERTa/checkpoint-5788/trainer_state.json +839 -0
  9. Internal-Run-EN/SpanBased-DeBERTa/checkpoint-5788/training_args.bin +3 -0
  10. Internal-Run-EN/SpanBased-DeBERTa/pytorch_model.bin +1 -1
Internal-Run-EN/SpanBased-DeBERTa/checkpoint-5788/config.json ADDED
@@ -0,0 +1,157 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "DebertaAdvancedSpanClassifier"
4
+ ],
5
+ "attention_probs_dropout_prob": 0.1,
6
+ "bos_token_id": null,
7
+ "dtype": "float32",
8
+ "eos_token_id": null,
9
+ "hidden_act": "gelu",
10
+ "hidden_dropout_prob": 0.1,
11
+ "hidden_size": 768,
12
+ "id2label": {
13
+ "0": "O",
14
+ "1": "ACCOUNTNAME",
15
+ "2": "ACCOUNTNUMBER",
16
+ "3": "AGE",
17
+ "4": "AMOUNT",
18
+ "5": "BIC",
19
+ "6": "BITCOINADDRESS",
20
+ "7": "BUILDINGNUMBER",
21
+ "8": "CITY",
22
+ "9": "COMPANYNAME",
23
+ "10": "COUNTY",
24
+ "11": "CREDITCARDCVV",
25
+ "12": "CREDITCARDISSUER",
26
+ "13": "CREDITCARDNUMBER",
27
+ "14": "CURRENCY",
28
+ "15": "CURRENCYCODE",
29
+ "16": "CURRENCYNAME",
30
+ "17": "CURRENCYSYMBOL",
31
+ "18": "DATE",
32
+ "19": "DOB",
33
+ "20": "EMAIL",
34
+ "21": "ETHEREUMADDRESS",
35
+ "22": "EYECOLOR",
36
+ "23": "FIRSTNAME",
37
+ "24": "GENDER",
38
+ "25": "HEIGHT",
39
+ "26": "IBAN",
40
+ "27": "IP",
41
+ "28": "IPV4",
42
+ "29": "IPV6",
43
+ "30": "JOBAREA",
44
+ "31": "JOBTITLE",
45
+ "32": "JOBTYPE",
46
+ "33": "LASTNAME",
47
+ "34": "LITECOINADDRESS",
48
+ "35": "MAC",
49
+ "36": "MASKEDNUMBER",
50
+ "37": "MIDDLENAME",
51
+ "38": "NEARBYGPSCOORDINATE",
52
+ "39": "ORDINALDIRECTION",
53
+ "40": "PASSWORD",
54
+ "41": "PHONEIMEI",
55
+ "42": "PHONENUMBER",
56
+ "43": "PIN",
57
+ "44": "PREFIX",
58
+ "45": "SECONDARYADDRESS",
59
+ "46": "SEX",
60
+ "47": "SSN",
61
+ "48": "STATE",
62
+ "49": "STREET",
63
+ "50": "TIME",
64
+ "51": "URL",
65
+ "52": "USERAGENT",
66
+ "53": "USERNAME",
67
+ "54": "VEHICLEVIN",
68
+ "55": "VEHICLEVRM",
69
+ "56": "ZIPCODE"
70
+ },
71
+ "initializer_range": 0.02,
72
+ "intermediate_size": 3072,
73
+ "label2id": {
74
+ "ACCOUNTNAME": 1,
75
+ "ACCOUNTNUMBER": 2,
76
+ "AGE": 3,
77
+ "AMOUNT": 4,
78
+ "BIC": 5,
79
+ "BITCOINADDRESS": 6,
80
+ "BUILDINGNUMBER": 7,
81
+ "CITY": 8,
82
+ "COMPANYNAME": 9,
83
+ "COUNTY": 10,
84
+ "CREDITCARDCVV": 11,
85
+ "CREDITCARDISSUER": 12,
86
+ "CREDITCARDNUMBER": 13,
87
+ "CURRENCY": 14,
88
+ "CURRENCYCODE": 15,
89
+ "CURRENCYNAME": 16,
90
+ "CURRENCYSYMBOL": 17,
91
+ "DATE": 18,
92
+ "DOB": 19,
93
+ "EMAIL": 20,
94
+ "ETHEREUMADDRESS": 21,
95
+ "EYECOLOR": 22,
96
+ "FIRSTNAME": 23,
97
+ "GENDER": 24,
98
+ "HEIGHT": 25,
99
+ "IBAN": 26,
100
+ "IP": 27,
101
+ "IPV4": 28,
102
+ "IPV6": 29,
103
+ "JOBAREA": 30,
104
+ "JOBTITLE": 31,
105
+ "JOBTYPE": 32,
106
+ "LASTNAME": 33,
107
+ "LITECOINADDRESS": 34,
108
+ "MAC": 35,
109
+ "MASKEDNUMBER": 36,
110
+ "MIDDLENAME": 37,
111
+ "NEARBYGPSCOORDINATE": 38,
112
+ "O": 0,
113
+ "ORDINALDIRECTION": 39,
114
+ "PASSWORD": 40,
115
+ "PHONEIMEI": 41,
116
+ "PHONENUMBER": 42,
117
+ "PIN": 43,
118
+ "PREFIX": 44,
119
+ "SECONDARYADDRESS": 45,
120
+ "SEX": 46,
121
+ "SSN": 47,
122
+ "STATE": 48,
123
+ "STREET": 49,
124
+ "TIME": 50,
125
+ "URL": 51,
126
+ "USERAGENT": 52,
127
+ "USERNAME": 53,
128
+ "VEHICLEVIN": 54,
129
+ "VEHICLEVRM": 55,
130
+ "ZIPCODE": 56
131
+ },
132
+ "layer_norm_eps": 1e-07,
133
+ "legacy": true,
134
+ "max_position_embeddings": 512,
135
+ "max_relative_positions": -1,
136
+ "model_type": "deberta-v2",
137
+ "norm_rel_ebd": "layer_norm",
138
+ "num_attention_heads": 12,
139
+ "num_hidden_layers": 12,
140
+ "pad_token_id": 0,
141
+ "pooler_dropout": 0,
142
+ "pooler_hidden_act": "gelu",
143
+ "pooler_hidden_size": 768,
144
+ "pos_att_type": [
145
+ "p2c",
146
+ "c2p"
147
+ ],
148
+ "position_biased_input": false,
149
+ "position_buckets": 256,
150
+ "relative_attention": true,
151
+ "share_att_key": true,
152
+ "tie_word_embeddings": true,
153
+ "transformers_version": "5.3.0",
154
+ "type_vocab_size": 0,
155
+ "use_cache": false,
156
+ "vocab_size": 128100
157
+ }
Internal-Run-EN/SpanBased-DeBERTa/checkpoint-5788/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:abc3b6c40ec506c814c98825bdb294f8f3bc482cb1827e061edc34a491e590a9
3
+ size 741504316
Internal-Run-EN/SpanBased-DeBERTa/checkpoint-5788/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:129e99f1df5d05d0af2ff4732a5fc15a28d01f66e7f97646a82dea91743cb73d
3
+ size 1483132299
Internal-Run-EN/SpanBased-DeBERTa/checkpoint-5788/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4f2d8a52ed2f3f158c09ed5ec3d81e9828b5d73cc6d7b249052fc29001c8fed0
3
+ size 14645
Internal-Run-EN/SpanBased-DeBERTa/checkpoint-5788/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:54ab1db15896f74728fbe94d1713ffbb2e68987979a89819bf869cf1209dd11f
3
+ size 1465
Internal-Run-EN/SpanBased-DeBERTa/checkpoint-5788/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
Internal-Run-EN/SpanBased-DeBERTa/checkpoint-5788/tokenizer_config.json ADDED
@@ -0,0 +1,23 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": true,
3
+ "backend": "tokenizers",
4
+ "bos_token": "[CLS]",
5
+ "cls_token": "[CLS]",
6
+ "do_lower_case": false,
7
+ "eos_token": "[SEP]",
8
+ "extra_special_tokens": [
9
+ "[PAD]",
10
+ "[CLS]",
11
+ "[SEP]"
12
+ ],
13
+ "is_local": false,
14
+ "mask_token": "[MASK]",
15
+ "model_max_length": 1000000000000000019884624838656,
16
+ "pad_token": "[PAD]",
17
+ "sep_token": "[SEP]",
18
+ "split_by_punct": false,
19
+ "tokenizer_class": "DebertaV2Tokenizer",
20
+ "unk_id": 3,
21
+ "unk_token": "[UNK]",
22
+ "vocab_type": "spm"
23
+ }
Internal-Run-EN/SpanBased-DeBERTa/checkpoint-5788/trainer_state.json ADDED
@@ -0,0 +1,839 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 2.0,
6
+ "eval_steps": 500,
7
+ "global_step": 5788,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.017277125086385625,
14
+ "grad_norm": 0.023754151538014412,
15
+ "learning_rate": 1.9830684174153422e-05,
16
+ "loss": 0.02889801502227783,
17
+ "step": 50
18
+ },
19
+ {
20
+ "epoch": 0.03455425017277125,
21
+ "grad_norm": 0.039325207471847534,
22
+ "learning_rate": 1.9657912923289565e-05,
23
+ "loss": 0.016851714849472045,
24
+ "step": 100
25
+ },
26
+ {
27
+ "epoch": 0.051831375259156875,
28
+ "grad_norm": 0.018957875669002533,
29
+ "learning_rate": 1.948514167242571e-05,
30
+ "loss": 0.014711350202560425,
31
+ "step": 150
32
+ },
33
+ {
34
+ "epoch": 0.0691085003455425,
35
+ "grad_norm": 0.02300567738711834,
36
+ "learning_rate": 1.9312370421561853e-05,
37
+ "loss": 0.013431484699249268,
38
+ "step": 200
39
+ },
40
+ {
41
+ "epoch": 0.08638562543192813,
42
+ "grad_norm": 0.018265269696712494,
43
+ "learning_rate": 1.9139599170698e-05,
44
+ "loss": 0.010455970764160155,
45
+ "step": 250
46
+ },
47
+ {
48
+ "epoch": 0.10366275051831375,
49
+ "grad_norm": 0.022894669324159622,
50
+ "learning_rate": 1.896682791983414e-05,
51
+ "loss": 0.008553104400634766,
52
+ "step": 300
53
+ },
54
+ {
55
+ "epoch": 0.12093987560469938,
56
+ "grad_norm": 0.02979782037436962,
57
+ "learning_rate": 1.8794056668970284e-05,
58
+ "loss": 0.006960532069206238,
59
+ "step": 350
60
+ },
61
+ {
62
+ "epoch": 0.138217000691085,
63
+ "grad_norm": 0.017046071588993073,
64
+ "learning_rate": 1.862128541810643e-05,
65
+ "loss": 0.006193249225616455,
66
+ "step": 400
67
+ },
68
+ {
69
+ "epoch": 0.15549412577747063,
70
+ "grad_norm": 0.014934674836695194,
71
+ "learning_rate": 1.8448514167242572e-05,
72
+ "loss": 0.005452051758766174,
73
+ "step": 450
74
+ },
75
+ {
76
+ "epoch": 0.17277125086385625,
77
+ "grad_norm": 0.018258241936564445,
78
+ "learning_rate": 1.8275742916378715e-05,
79
+ "loss": 0.004681007862091064,
80
+ "step": 500
81
+ },
82
+ {
83
+ "epoch": 0.19004837595024188,
84
+ "grad_norm": 0.022132372483611107,
85
+ "learning_rate": 1.810297166551486e-05,
86
+ "loss": 0.004303434491157531,
87
+ "step": 550
88
+ },
89
+ {
90
+ "epoch": 0.2073255010366275,
91
+ "grad_norm": 0.025860296562314034,
92
+ "learning_rate": 1.7930200414651003e-05,
93
+ "loss": 0.003719818592071533,
94
+ "step": 600
95
+ },
96
+ {
97
+ "epoch": 0.22460262612301313,
98
+ "grad_norm": 0.01564023829996586,
99
+ "learning_rate": 1.775742916378715e-05,
100
+ "loss": 0.0033052828907966614,
101
+ "step": 650
102
+ },
103
+ {
104
+ "epoch": 0.24187975120939875,
105
+ "grad_norm": 0.014180956408381462,
106
+ "learning_rate": 1.758465791292329e-05,
107
+ "loss": 0.0030368798971176147,
108
+ "step": 700
109
+ },
110
+ {
111
+ "epoch": 0.2591568762957844,
112
+ "grad_norm": 0.024306217208504677,
113
+ "learning_rate": 1.7411886662059434e-05,
114
+ "loss": 0.0027125704288482664,
115
+ "step": 750
116
+ },
117
+ {
118
+ "epoch": 0.27643400138217,
119
+ "grad_norm": 0.017899125814437866,
120
+ "learning_rate": 1.723911541119558e-05,
121
+ "loss": 0.0024478091299533844,
122
+ "step": 800
123
+ },
124
+ {
125
+ "epoch": 0.29371112646855563,
126
+ "grad_norm": 0.012664858251810074,
127
+ "learning_rate": 1.7066344160331722e-05,
128
+ "loss": 0.002277335822582245,
129
+ "step": 850
130
+ },
131
+ {
132
+ "epoch": 0.31098825155494125,
133
+ "grad_norm": 0.01861223205924034,
134
+ "learning_rate": 1.6893572909467865e-05,
135
+ "loss": 0.0020104184746742248,
136
+ "step": 900
137
+ },
138
+ {
139
+ "epoch": 0.3282653766413269,
140
+ "grad_norm": 0.01881146989762783,
141
+ "learning_rate": 1.6720801658604007e-05,
142
+ "loss": 0.0017749948799610138,
143
+ "step": 950
144
+ },
145
+ {
146
+ "epoch": 0.3455425017277125,
147
+ "grad_norm": 0.016596628352999687,
148
+ "learning_rate": 1.6548030407740153e-05,
149
+ "loss": 0.0017245493829250336,
150
+ "step": 1000
151
+ },
152
+ {
153
+ "epoch": 0.36281962681409813,
154
+ "grad_norm": 0.019588826224207878,
155
+ "learning_rate": 1.63752591568763e-05,
156
+ "loss": 0.0016060712933540344,
157
+ "step": 1050
158
+ },
159
+ {
160
+ "epoch": 0.38009675190048375,
161
+ "grad_norm": 0.01674879714846611,
162
+ "learning_rate": 1.620248790601244e-05,
163
+ "loss": 0.0015378570556640624,
164
+ "step": 1100
165
+ },
166
+ {
167
+ "epoch": 0.3973738769868694,
168
+ "grad_norm": 0.013815218582749367,
169
+ "learning_rate": 1.6029716655148584e-05,
170
+ "loss": 0.0014464271068572998,
171
+ "step": 1150
172
+ },
173
+ {
174
+ "epoch": 0.414651002073255,
175
+ "grad_norm": 0.010738871991634369,
176
+ "learning_rate": 1.585694540428473e-05,
177
+ "loss": 0.0013705722987651825,
178
+ "step": 1200
179
+ },
180
+ {
181
+ "epoch": 0.43192812715964063,
182
+ "grad_norm": 0.010381566360592842,
183
+ "learning_rate": 1.5684174153420872e-05,
184
+ "loss": 0.0012675783038139344,
185
+ "step": 1250
186
+ },
187
+ {
188
+ "epoch": 0.44920525224602625,
189
+ "grad_norm": 0.011795097030699253,
190
+ "learning_rate": 1.5511402902557015e-05,
191
+ "loss": 0.001222059577703476,
192
+ "step": 1300
193
+ },
194
+ {
195
+ "epoch": 0.4664823773324119,
196
+ "grad_norm": 0.014987198635935783,
197
+ "learning_rate": 1.5338631651693157e-05,
198
+ "loss": 0.001066400781273842,
199
+ "step": 1350
200
+ },
201
+ {
202
+ "epoch": 0.4837595024187975,
203
+ "grad_norm": 0.014591517858207226,
204
+ "learning_rate": 1.5165860400829305e-05,
205
+ "loss": 0.0010450337082147597,
206
+ "step": 1400
207
+ },
208
+ {
209
+ "epoch": 0.5010366275051832,
210
+ "grad_norm": 0.01359975803643465,
211
+ "learning_rate": 1.4993089149965447e-05,
212
+ "loss": 0.0010743821412324906,
213
+ "step": 1450
214
+ },
215
+ {
216
+ "epoch": 0.5183137525915688,
217
+ "grad_norm": 0.021828675642609596,
218
+ "learning_rate": 1.482031789910159e-05,
219
+ "loss": 0.0009260432422161102,
220
+ "step": 1500
221
+ },
222
+ {
223
+ "epoch": 0.5355908776779544,
224
+ "grad_norm": 0.012535043992102146,
225
+ "learning_rate": 1.4647546648237734e-05,
226
+ "loss": 0.0009216105937957763,
227
+ "step": 1550
228
+ },
229
+ {
230
+ "epoch": 0.55286800276434,
231
+ "grad_norm": 0.019430212676525116,
232
+ "learning_rate": 1.447477539737388e-05,
233
+ "loss": 0.000878768116235733,
234
+ "step": 1600
235
+ },
236
+ {
237
+ "epoch": 0.5701451278507257,
238
+ "grad_norm": 0.017275096848607063,
239
+ "learning_rate": 1.4302004146510022e-05,
240
+ "loss": 0.000844210758805275,
241
+ "step": 1650
242
+ },
243
+ {
244
+ "epoch": 0.5874222529371113,
245
+ "grad_norm": 0.01682225428521633,
246
+ "learning_rate": 1.4129232895646165e-05,
247
+ "loss": 0.000815606415271759,
248
+ "step": 1700
249
+ },
250
+ {
251
+ "epoch": 0.6046993780234969,
252
+ "grad_norm": 0.009485620073974133,
253
+ "learning_rate": 1.3956461644782309e-05,
254
+ "loss": 0.0007856705039739609,
255
+ "step": 1750
256
+ },
257
+ {
258
+ "epoch": 0.6219765031098825,
259
+ "grad_norm": 0.029362259432673454,
260
+ "learning_rate": 1.3783690393918455e-05,
261
+ "loss": 0.0007757318764925003,
262
+ "step": 1800
263
+ },
264
+ {
265
+ "epoch": 0.6392536281962682,
266
+ "grad_norm": 0.012905504554510117,
267
+ "learning_rate": 1.3610919143054597e-05,
268
+ "loss": 0.0007183963060379029,
269
+ "step": 1850
270
+ },
271
+ {
272
+ "epoch": 0.6565307532826538,
273
+ "grad_norm": 0.007750813849270344,
274
+ "learning_rate": 1.343814789219074e-05,
275
+ "loss": 0.0007490935921669006,
276
+ "step": 1900
277
+ },
278
+ {
279
+ "epoch": 0.6738078783690394,
280
+ "grad_norm": 0.016771988943219185,
281
+ "learning_rate": 1.3265376641326884e-05,
282
+ "loss": 0.0007053028047084808,
283
+ "step": 1950
284
+ },
285
+ {
286
+ "epoch": 0.691085003455425,
287
+ "grad_norm": 0.013029007241129875,
288
+ "learning_rate": 1.309260539046303e-05,
289
+ "loss": 0.0006876800209283829,
290
+ "step": 2000
291
+ },
292
+ {
293
+ "epoch": 0.7083621285418107,
294
+ "grad_norm": 0.01617145538330078,
295
+ "learning_rate": 1.2919834139599172e-05,
296
+ "loss": 0.0006662622094154358,
297
+ "step": 2050
298
+ },
299
+ {
300
+ "epoch": 0.7256392536281963,
301
+ "grad_norm": 0.0077132186852395535,
302
+ "learning_rate": 1.2747062888735315e-05,
303
+ "loss": 0.0006389703601598739,
304
+ "step": 2100
305
+ },
306
+ {
307
+ "epoch": 0.7429163787145819,
308
+ "grad_norm": 0.01814083755016327,
309
+ "learning_rate": 1.2574291637871459e-05,
310
+ "loss": 0.0006544689834117889,
311
+ "step": 2150
312
+ },
313
+ {
314
+ "epoch": 0.7601935038009675,
315
+ "grad_norm": 0.009747780859470367,
316
+ "learning_rate": 1.2401520387007605e-05,
317
+ "loss": 0.0006402868777513504,
318
+ "step": 2200
319
+ },
320
+ {
321
+ "epoch": 0.7774706288873532,
322
+ "grad_norm": 0.014642719179391861,
323
+ "learning_rate": 1.2228749136143747e-05,
324
+ "loss": 0.0005662843957543373,
325
+ "step": 2250
326
+ },
327
+ {
328
+ "epoch": 0.7947477539737388,
329
+ "grad_norm": 0.013663463294506073,
330
+ "learning_rate": 1.205597788527989e-05,
331
+ "loss": 0.0006011456251144409,
332
+ "step": 2300
333
+ },
334
+ {
335
+ "epoch": 0.8120248790601244,
336
+ "grad_norm": 0.017723126336932182,
337
+ "learning_rate": 1.1883206634416034e-05,
338
+ "loss": 0.000600893571972847,
339
+ "step": 2350
340
+ },
341
+ {
342
+ "epoch": 0.82930200414651,
343
+ "grad_norm": 0.018065618351101875,
344
+ "learning_rate": 1.1710435383552178e-05,
345
+ "loss": 0.0005438444763422013,
346
+ "step": 2400
347
+ },
348
+ {
349
+ "epoch": 0.8465791292328957,
350
+ "grad_norm": 0.010403930209577084,
351
+ "learning_rate": 1.1537664132688322e-05,
352
+ "loss": 0.0005565705522894859,
353
+ "step": 2450
354
+ },
355
+ {
356
+ "epoch": 0.8638562543192813,
357
+ "grad_norm": 0.01131166610866785,
358
+ "learning_rate": 1.1364892881824465e-05,
359
+ "loss": 0.0005076530575752258,
360
+ "step": 2500
361
+ },
362
+ {
363
+ "epoch": 0.8811333794056669,
364
+ "grad_norm": 0.01727471500635147,
365
+ "learning_rate": 1.1192121630960609e-05,
366
+ "loss": 0.0005273338034749031,
367
+ "step": 2550
368
+ },
369
+ {
370
+ "epoch": 0.8984105044920525,
371
+ "grad_norm": 0.009765254333615303,
372
+ "learning_rate": 1.1019350380096753e-05,
373
+ "loss": 0.0004916878044605255,
374
+ "step": 2600
375
+ },
376
+ {
377
+ "epoch": 0.9156876295784382,
378
+ "grad_norm": 0.009817619808018208,
379
+ "learning_rate": 1.0846579129232897e-05,
380
+ "loss": 0.00046188205480575563,
381
+ "step": 2650
382
+ },
383
+ {
384
+ "epoch": 0.9329647546648238,
385
+ "grad_norm": 0.012648608535528183,
386
+ "learning_rate": 1.067380787836904e-05,
387
+ "loss": 0.0005124472826719284,
388
+ "step": 2700
389
+ },
390
+ {
391
+ "epoch": 0.9502418797512094,
392
+ "grad_norm": 0.00930147897452116,
393
+ "learning_rate": 1.0501036627505184e-05,
394
+ "loss": 0.0005098021402955055,
395
+ "step": 2750
396
+ },
397
+ {
398
+ "epoch": 0.967519004837595,
399
+ "grad_norm": 0.008392088115215302,
400
+ "learning_rate": 1.0328265376641328e-05,
401
+ "loss": 0.0004884525388479233,
402
+ "step": 2800
403
+ },
404
+ {
405
+ "epoch": 0.9847961299239807,
406
+ "grad_norm": 0.008711796253919601,
407
+ "learning_rate": 1.0155494125777472e-05,
408
+ "loss": 0.00047579750418663024,
409
+ "step": 2850
410
+ },
411
+ {
412
+ "epoch": 1.0020732550103664,
413
+ "grad_norm": 0.012193967588245869,
414
+ "learning_rate": 9.982722874913615e-06,
415
+ "loss": 0.0004131799191236496,
416
+ "step": 2900
417
+ },
418
+ {
419
+ "epoch": 1.0193503800967518,
420
+ "grad_norm": 0.009380247443914413,
421
+ "learning_rate": 9.809951624049759e-06,
422
+ "loss": 0.00042345020920038224,
423
+ "step": 2950
424
+ },
425
+ {
426
+ "epoch": 1.0366275051831375,
427
+ "grad_norm": 0.00867867935448885,
428
+ "learning_rate": 9.637180373185903e-06,
429
+ "loss": 0.0003872714564204216,
430
+ "step": 3000
431
+ },
432
+ {
433
+ "epoch": 1.0539046302695232,
434
+ "grad_norm": 0.007016318384557962,
435
+ "learning_rate": 9.464409122322047e-06,
436
+ "loss": 0.0003861965239048004,
437
+ "step": 3050
438
+ },
439
+ {
440
+ "epoch": 1.0711817553559089,
441
+ "grad_norm": 0.004630661569535732,
442
+ "learning_rate": 9.29163787145819e-06,
443
+ "loss": 0.0003987585380673408,
444
+ "step": 3100
445
+ },
446
+ {
447
+ "epoch": 1.0884588804422943,
448
+ "grad_norm": 0.006467690225690603,
449
+ "learning_rate": 9.118866620594334e-06,
450
+ "loss": 0.00040752418339252473,
451
+ "step": 3150
452
+ },
453
+ {
454
+ "epoch": 1.10573600552868,
455
+ "grad_norm": 0.005555268377065659,
456
+ "learning_rate": 8.946095369730478e-06,
457
+ "loss": 0.0003801156580448151,
458
+ "step": 3200
459
+ },
460
+ {
461
+ "epoch": 1.1230131306150657,
462
+ "grad_norm": 0.004987559746950865,
463
+ "learning_rate": 8.773324118866622e-06,
464
+ "loss": 0.0003666750341653824,
465
+ "step": 3250
466
+ },
467
+ {
468
+ "epoch": 1.1402902557014514,
469
+ "grad_norm": 0.008231359533965588,
470
+ "learning_rate": 8.600552868002765e-06,
471
+ "loss": 0.0004398062825202942,
472
+ "step": 3300
473
+ },
474
+ {
475
+ "epoch": 1.1575673807878368,
476
+ "grad_norm": 0.0057737682946026325,
477
+ "learning_rate": 8.427781617138909e-06,
478
+ "loss": 0.00038419820368289947,
479
+ "step": 3350
480
+ },
481
+ {
482
+ "epoch": 1.1748445058742225,
483
+ "grad_norm": 0.006889602169394493,
484
+ "learning_rate": 8.255010366275053e-06,
485
+ "loss": 0.00039252057671546937,
486
+ "step": 3400
487
+ },
488
+ {
489
+ "epoch": 1.1921216309606082,
490
+ "grad_norm": 0.01179490052163601,
491
+ "learning_rate": 8.082239115411197e-06,
492
+ "loss": 0.0003788643702864647,
493
+ "step": 3450
494
+ },
495
+ {
496
+ "epoch": 1.2093987560469939,
497
+ "grad_norm": 0.008482358418405056,
498
+ "learning_rate": 7.90946786454734e-06,
499
+ "loss": 0.00035284604877233505,
500
+ "step": 3500
501
+ },
502
+ {
503
+ "epoch": 1.2266758811333793,
504
+ "grad_norm": 0.009830164723098278,
505
+ "learning_rate": 7.736696613683484e-06,
506
+ "loss": 0.0003476132825016975,
507
+ "step": 3550
508
+ },
509
+ {
510
+ "epoch": 1.243953006219765,
511
+ "grad_norm": 0.006578364875167608,
512
+ "learning_rate": 7.563925362819627e-06,
513
+ "loss": 0.00034327574074268343,
514
+ "step": 3600
515
+ },
516
+ {
517
+ "epoch": 1.2612301313061507,
518
+ "grad_norm": 0.00881590973585844,
519
+ "learning_rate": 7.391154111955771e-06,
520
+ "loss": 0.00037471447139978407,
521
+ "step": 3650
522
+ },
523
+ {
524
+ "epoch": 1.2785072563925364,
525
+ "grad_norm": 0.004738886374980211,
526
+ "learning_rate": 7.218382861091915e-06,
527
+ "loss": 0.00032851655036211016,
528
+ "step": 3700
529
+ },
530
+ {
531
+ "epoch": 1.2957843814789218,
532
+ "grad_norm": 0.004191044718027115,
533
+ "learning_rate": 7.045611610228059e-06,
534
+ "loss": 0.00032500557601451875,
535
+ "step": 3750
536
+ },
537
+ {
538
+ "epoch": 1.3130615065653075,
539
+ "grad_norm": 0.007185476366430521,
540
+ "learning_rate": 6.872840359364202e-06,
541
+ "loss": 0.000325874425470829,
542
+ "step": 3800
543
+ },
544
+ {
545
+ "epoch": 1.3303386316516932,
546
+ "grad_norm": 0.010816428810358047,
547
+ "learning_rate": 6.700069108500346e-06,
548
+ "loss": 0.000360981710255146,
549
+ "step": 3850
550
+ },
551
+ {
552
+ "epoch": 1.3476157567380787,
553
+ "grad_norm": 0.0067063430324196815,
554
+ "learning_rate": 6.52729785763649e-06,
555
+ "loss": 0.0003418443351984024,
556
+ "step": 3900
557
+ },
558
+ {
559
+ "epoch": 1.3648928818244643,
560
+ "grad_norm": 0.010665838606655598,
561
+ "learning_rate": 6.354526606772633e-06,
562
+ "loss": 0.0003797527030110359,
563
+ "step": 3950
564
+ },
565
+ {
566
+ "epoch": 1.38217000691085,
567
+ "grad_norm": 0.006655356381088495,
568
+ "learning_rate": 6.181755355908777e-06,
569
+ "loss": 0.0003413737565279007,
570
+ "step": 4000
571
+ },
572
+ {
573
+ "epoch": 1.3994471319972357,
574
+ "grad_norm": 0.008293711580336094,
575
+ "learning_rate": 6.0089841050449205e-06,
576
+ "loss": 0.0003274304419755936,
577
+ "step": 4050
578
+ },
579
+ {
580
+ "epoch": 1.4167242570836214,
581
+ "grad_norm": 0.015750911086797714,
582
+ "learning_rate": 5.836212854181065e-06,
583
+ "loss": 0.00028720928356051445,
584
+ "step": 4100
585
+ },
586
+ {
587
+ "epoch": 1.4340013821700068,
588
+ "grad_norm": 0.008319740183651447,
589
+ "learning_rate": 5.663441603317208e-06,
590
+ "loss": 0.0003450794890522957,
591
+ "step": 4150
592
+ },
593
+ {
594
+ "epoch": 1.4512785072563925,
595
+ "grad_norm": 0.004470289219170809,
596
+ "learning_rate": 5.490670352453352e-06,
597
+ "loss": 0.00031472038477659226,
598
+ "step": 4200
599
+ },
600
+ {
601
+ "epoch": 1.4685556323427782,
602
+ "grad_norm": 0.006401464808732271,
603
+ "learning_rate": 5.3178991015894955e-06,
604
+ "loss": 0.00030458766967058183,
605
+ "step": 4250
606
+ },
607
+ {
608
+ "epoch": 1.4858327574291637,
609
+ "grad_norm": 0.010065151378512383,
610
+ "learning_rate": 5.14512785072564e-06,
611
+ "loss": 0.0003296390920877457,
612
+ "step": 4300
613
+ },
614
+ {
615
+ "epoch": 1.5031098825155493,
616
+ "grad_norm": 0.00851607508957386,
617
+ "learning_rate": 4.972356599861784e-06,
618
+ "loss": 0.0003387167677283287,
619
+ "step": 4350
620
+ },
621
+ {
622
+ "epoch": 1.520387007601935,
623
+ "grad_norm": 0.00729724857956171,
624
+ "learning_rate": 4.799585348997927e-06,
625
+ "loss": 0.00032126672565937044,
626
+ "step": 4400
627
+ },
628
+ {
629
+ "epoch": 1.5376641326883207,
630
+ "grad_norm": 0.006632724776864052,
631
+ "learning_rate": 4.6268140981340705e-06,
632
+ "loss": 0.00031396977603435515,
633
+ "step": 4450
634
+ },
635
+ {
636
+ "epoch": 1.5549412577747064,
637
+ "grad_norm": 0.00762026896700263,
638
+ "learning_rate": 4.454042847270215e-06,
639
+ "loss": 0.0003191409632563591,
640
+ "step": 4500
641
+ },
642
+ {
643
+ "epoch": 1.572218382861092,
644
+ "grad_norm": 0.010211093351244926,
645
+ "learning_rate": 4.281271596406358e-06,
646
+ "loss": 0.00031375911086797717,
647
+ "step": 4550
648
+ },
649
+ {
650
+ "epoch": 1.5894955079474775,
651
+ "grad_norm": 0.007663295604288578,
652
+ "learning_rate": 4.108500345542502e-06,
653
+ "loss": 0.0002942489646375179,
654
+ "step": 4600
655
+ },
656
+ {
657
+ "epoch": 1.6067726330338632,
658
+ "grad_norm": 0.016654301434755325,
659
+ "learning_rate": 3.9357290946786455e-06,
660
+ "loss": 0.0003081130795180798,
661
+ "step": 4650
662
+ },
663
+ {
664
+ "epoch": 1.6240497581202487,
665
+ "grad_norm": 0.007844490930438042,
666
+ "learning_rate": 3.7629578438147897e-06,
667
+ "loss": 0.00028737472370266915,
668
+ "step": 4700
669
+ },
670
+ {
671
+ "epoch": 1.6413268832066343,
672
+ "grad_norm": 0.004168346058577299,
673
+ "learning_rate": 3.5901865929509334e-06,
674
+ "loss": 0.00026899419724941254,
675
+ "step": 4750
676
+ },
677
+ {
678
+ "epoch": 1.65860400829302,
679
+ "grad_norm": 0.005883086938410997,
680
+ "learning_rate": 3.417415342087077e-06,
681
+ "loss": 0.0003354706987738609,
682
+ "step": 4800
683
+ },
684
+ {
685
+ "epoch": 1.6758811333794057,
686
+ "grad_norm": 0.0029302723705768585,
687
+ "learning_rate": 3.244644091223221e-06,
688
+ "loss": 0.00030305808410048486,
689
+ "step": 4850
690
+ },
691
+ {
692
+ "epoch": 1.6931582584657914,
693
+ "grad_norm": 0.006633299868553877,
694
+ "learning_rate": 3.0718728403593647e-06,
695
+ "loss": 0.0002642534300684929,
696
+ "step": 4900
697
+ },
698
+ {
699
+ "epoch": 1.710435383552177,
700
+ "grad_norm": 0.008219488896429539,
701
+ "learning_rate": 2.8991015894955084e-06,
702
+ "loss": 0.00029796602204442026,
703
+ "step": 4950
704
+ },
705
+ {
706
+ "epoch": 1.7277125086385625,
707
+ "grad_norm": 0.012344489805400372,
708
+ "learning_rate": 2.726330338631652e-06,
709
+ "loss": 0.00026235152035951613,
710
+ "step": 5000
711
+ },
712
+ {
713
+ "epoch": 1.7449896337249482,
714
+ "grad_norm": 0.007912565022706985,
715
+ "learning_rate": 2.553559087767796e-06,
716
+ "loss": 0.0002975315414369106,
717
+ "step": 5050
718
+ },
719
+ {
720
+ "epoch": 1.7622667588113337,
721
+ "grad_norm": 0.010620102286338806,
722
+ "learning_rate": 2.3807878369039396e-06,
723
+ "loss": 0.0003033735789358616,
724
+ "step": 5100
725
+ },
726
+ {
727
+ "epoch": 1.7795438838977193,
728
+ "grad_norm": 0.0051605780608952045,
729
+ "learning_rate": 2.208016586040083e-06,
730
+ "loss": 0.00029248302802443505,
731
+ "step": 5150
732
+ },
733
+ {
734
+ "epoch": 1.796821008984105,
735
+ "grad_norm": 0.006801155861467123,
736
+ "learning_rate": 2.0352453351762267e-06,
737
+ "loss": 0.0002811450697481632,
738
+ "step": 5200
739
+ },
740
+ {
741
+ "epoch": 1.8140981340704907,
742
+ "grad_norm": 0.008665273897349834,
743
+ "learning_rate": 1.8624740843123707e-06,
744
+ "loss": 0.0002777145616710186,
745
+ "step": 5250
746
+ },
747
+ {
748
+ "epoch": 1.8313752591568764,
749
+ "grad_norm": 0.007553344126790762,
750
+ "learning_rate": 1.6897028334485144e-06,
751
+ "loss": 0.0002958022244274616,
752
+ "step": 5300
753
+ },
754
+ {
755
+ "epoch": 1.848652384243262,
756
+ "grad_norm": 0.005073766689747572,
757
+ "learning_rate": 1.5169315825846582e-06,
758
+ "loss": 0.00026641294360160825,
759
+ "step": 5350
760
+ },
761
+ {
762
+ "epoch": 1.8659295093296475,
763
+ "grad_norm": 0.00378187233582139,
764
+ "learning_rate": 1.344160331720802e-06,
765
+ "loss": 0.00028496159240603446,
766
+ "step": 5400
767
+ },
768
+ {
769
+ "epoch": 1.8832066344160332,
770
+ "grad_norm": 0.00831520464271307,
771
+ "learning_rate": 1.1713890808569455e-06,
772
+ "loss": 0.0002626798860728741,
773
+ "step": 5450
774
+ },
775
+ {
776
+ "epoch": 1.9004837595024187,
777
+ "grad_norm": 0.00719051668420434,
778
+ "learning_rate": 9.986178299930892e-07,
779
+ "loss": 0.0002551780082285404,
780
+ "step": 5500
781
+ },
782
+ {
783
+ "epoch": 1.9177608845888043,
784
+ "grad_norm": 0.006325442809611559,
785
+ "learning_rate": 8.25846579129233e-07,
786
+ "loss": 0.0002771000005304813,
787
+ "step": 5550
788
+ },
789
+ {
790
+ "epoch": 1.93503800967519,
791
+ "grad_norm": 0.013826875947415829,
792
+ "learning_rate": 6.530753282653766e-07,
793
+ "loss": 0.00027303243055939677,
794
+ "step": 5600
795
+ },
796
+ {
797
+ "epoch": 1.9523151347615757,
798
+ "grad_norm": 0.010165671817958355,
799
+ "learning_rate": 4.803040774015205e-07,
800
+ "loss": 0.0002723958715796471,
801
+ "step": 5650
802
+ },
803
+ {
804
+ "epoch": 1.9695922598479614,
805
+ "grad_norm": 0.008166917599737644,
806
+ "learning_rate": 3.0753282653766415e-07,
807
+ "loss": 0.000264077503234148,
808
+ "step": 5700
809
+ },
810
+ {
811
+ "epoch": 1.986869384934347,
812
+ "grad_norm": 0.0023579970002174377,
813
+ "learning_rate": 1.3476157567380787e-07,
814
+ "loss": 0.0002627692371606827,
815
+ "step": 5750
816
+ }
817
+ ],
818
+ "logging_steps": 50,
819
+ "max_steps": 5788,
820
+ "num_input_tokens_seen": 0,
821
+ "num_train_epochs": 2,
822
+ "save_steps": 500,
823
+ "stateful_callbacks": {
824
+ "TrainerControl": {
825
+ "args": {
826
+ "should_epoch_stop": false,
827
+ "should_evaluate": false,
828
+ "should_log": false,
829
+ "should_save": true,
830
+ "should_training_stop": true
831
+ },
832
+ "attributes": {}
833
+ }
834
+ },
835
+ "total_flos": 1.818607033149919e+16,
836
+ "train_batch_size": 64,
837
+ "trial_name": null,
838
+ "trial_params": null
839
+ }
Internal-Run-EN/SpanBased-DeBERTa/checkpoint-5788/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:65050eab37d8ac961760da005e0bfc5559c3d1c3f07bcb3c26dc7d3020ec6fe1
3
+ size 5201
Internal-Run-EN/SpanBased-DeBERTa/pytorch_model.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:7c139ce6e098c2a976524ea68bc2e666b12d1c00c3d4d157ba7bc0e43574e829
3
  size 741566223
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4bc3e764d3cd8ffd9c75481ea14c96d6bd0eec3adb737de22bcff5f2d21e59af
3
  size 741566223