Hugging Face's logo

deprem-ml
/
name_anonymization

Token Classification
Transformers
PyTorch
Turkish
bert
ner
Model card Files
xet
 Community
1
akoksal commited on
Commit
e724741
·
1 Parent(s): 0d33a10

Upload with huggingface_hub

Browse files
Files changed (11) hide show
  1. config.json +71 -0
  2. optimizer.pt +3 -0
  3. pytorch_model.bin +3 -0
  4. rng_state.pth +3 -0
  5. scheduler.pt +3 -0
  6. special_tokens_map.json +7 -0
  7. tokenizer.json +0 -0
  8. tokenizer_config.json +17 -0
  9. trainer_state.json +306 -0
  10. training_args.bin +3 -0
  11. vocab.txt +0 -0
config.json ADDED
@@ -0,0 +1,71 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "dbmdz/bert-base-turkish-cased",
3
+ "architectures": [
4
+ "BertForTokenClassification"
5
+ ],
6
+ "attention_probs_dropout_prob": 0.1,
7
+ "classifier_dropout": null,
8
+ "hidden_act": "gelu",
9
+ "hidden_dropout_prob": 0.1,
10
+ "hidden_size": 768,
11
+ "id2label": {
12
+ "0": "O",
13
+ "1": "B-no",
14
+ "2": "I-no",
15
+ "3": "B-mahalle",
16
+ "4": "I-mahalle",
17
+ "5": "B-ad-soyad",
18
+ "6": "I-ad-soyad",
19
+ "7": "B-kat",
20
+ "8": "I-kat",
21
+ "9": "B-ilce",
22
+ "10": "I-ilce",
23
+ "11": "B-sokak",
24
+ "12": "I-sokak",
25
+ "13": "B-dis kapi no",
26
+ "14": "I-dis kapi no",
27
+ "15": "B-Apartman/Site",
28
+ "16": "I-Apartman/Site",
29
+ "17": "B-ic kapi no",
30
+ "18": "I-ic kapi no",
31
+ "19": "B-il",
32
+ "20": "I-il"
33
+ },
34
+ "initializer_range": 0.02,
35
+ "intermediate_size": 3072,
36
+ "label2id": {
37
+ "LABEL_0": 0,
38
+ "LABEL_1": 1,
39
+ "LABEL_10": 10,
40
+ "LABEL_11": 11,
41
+ "LABEL_12": 12,
42
+ "LABEL_13": 13,
43
+ "LABEL_14": 14,
44
+ "LABEL_15": 15,
45
+ "LABEL_16": 16,
46
+ "LABEL_17": 17,
47
+ "LABEL_18": 18,
48
+ "LABEL_19": 19,
49
+ "LABEL_2": 2,
50
+ "LABEL_20": 20,
51
+ "LABEL_3": 3,
52
+ "LABEL_4": 4,
53
+ "LABEL_5": 5,
54
+ "LABEL_6": 6,
55
+ "LABEL_7": 7,
56
+ "LABEL_8": 8,
57
+ "LABEL_9": 9
58
+ },
59
+ "layer_norm_eps": 1e-12,
60
+ "max_position_embeddings": 512,
61
+ "model_type": "bert",
62
+ "num_attention_heads": 12,
63
+ "num_hidden_layers": 12,
64
+ "pad_token_id": 0,
65
+ "position_embedding_type": "absolute",
66
+ "torch_dtype": "float32",
67
+ "transformers_version": "4.26.1",
68
+ "type_vocab_size": 2,
69
+ "use_cache": true,
70
+ "vocab_size": 32000
71
+ }
optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b1b218ede1e774f4ee70f7198dff9ca381914a7584b74cb3e3ca558b902d3d47
3
+ size 880458405
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d232bd94b71cf976b86e405f66a9977479954327e95c47d07c53c741cee057ad
3
+ size 440242033
rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:978b385d0eac0bd747225a9767580bbc0718dd5abca4bb7fef9d4fd5fc49171e
3
+ size 14503
scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f79e9e7e0cc7d63fe68db3376b63b0f58f3124e24cf6242343661b597c4f9afe
3
+ size 623
special_tokens_map.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "cls_token": "[CLS]",
3
+ "mask_token": "[MASK]",
4
+ "pad_token": "[PAD]",
5
+ "sep_token": "[SEP]",
6
+ "unk_token": "[UNK]"
7
+ }
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer_config.json ADDED
@@ -0,0 +1,17 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "cls_token": "[CLS]",
3
+ "do_basic_tokenize": true,
4
+ "do_lower_case": false,
5
+ "mask_token": "[MASK]",
6
+ "max_len": 512,
7
+ "model_max_length": 512,
8
+ "name_or_path": "dbmdz/bert-base-turkish-cased",
9
+ "never_split": null,
10
+ "pad_token": "[PAD]",
11
+ "sep_token": "[SEP]",
12
+ "special_tokens_map_file": null,
13
+ "strip_accents": null,
14
+ "tokenize_chinese_chars": true,
15
+ "tokenizer_class": "BertTokenizer",
16
+ "unk_token": "[UNK]"
17
+ }
trainer_state.json ADDED
@@ -0,0 +1,306 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.252394437789917,
3
+ "best_model_checkpoint": "/mounts/work/akoksal/earthquake_ner_models/checkpoint-192",
4
+ "epoch": 5.0,
5
+ "global_step": 240,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 1.0,
12
+ "learning_rate": 4.4444444444444447e-05,
13
+ "loss": 1.1871,
14
+ "step": 48
15
+ },
16
+ {
17
+ "epoch": 1.0,
18
+ "eval_Apartman/Site_f1": 0.4812834224598931,
19
+ "eval_Apartman/Site_precision": 0.4017857142857143,
20
+ "eval_Apartman/Site_recall": 0.6,
21
+ "eval_Apartman/Site_support": 75,
22
+ "eval_ad-soyad_f1": 0.7657992565055761,
23
+ "eval_ad-soyad_precision": 0.7006802721088435,
24
+ "eval_ad-soyad_recall": 0.8442622950819673,
25
+ "eval_ad-soyad_support": 122,
26
+ "eval_dis kapi no_f1": 0.0,
27
+ "eval_dis kapi no_precision": 0.0,
28
+ "eval_dis kapi no_recall": 0.0,
29
+ "eval_dis kapi no_support": 9,
30
+ "eval_ic kapi no_f1": 0.0,
31
+ "eval_ic kapi no_precision": 0.0,
32
+ "eval_ic kapi no_recall": 0.0,
33
+ "eval_ic kapi no_support": 10,
34
+ "eval_il_f1": 0.7532467532467533,
35
+ "eval_il_precision": 0.696,
36
+ "eval_il_recall": 0.8207547169811321,
37
+ "eval_il_support": 106,
38
+ "eval_ilce_f1": 0.5732484076433121,
39
+ "eval_ilce_precision": 0.6164383561643836,
40
+ "eval_ilce_recall": 0.5357142857142857,
41
+ "eval_ilce_support": 84,
42
+ "eval_kat_f1": 0.0,
43
+ "eval_kat_precision": 0.0,
44
+ "eval_kat_recall": 0.0,
45
+ "eval_kat_support": 13,
46
+ "eval_loss": 0.34245046973228455,
47
+ "eval_mahalle_f1": 0.6117647058823529,
48
+ "eval_mahalle_precision": 0.5454545454545454,
49
+ "eval_mahalle_recall": 0.6964285714285714,
50
+ "eval_mahalle_support": 112,
51
+ "eval_no_f1": 0.33557046979865773,
52
+ "eval_no_precision": 0.2336448598130841,
53
+ "eval_no_recall": 0.5952380952380952,
54
+ "eval_no_support": 42,
55
+ "eval_overall_accuracy": 0.9031231350706187,
56
+ "eval_overall_f1": 0.5896103896103896,
57
+ "eval_overall_precision": 0.530373831775701,
58
+ "eval_overall_recall": 0.6637426900584795,
59
+ "eval_runtime": 0.5655,
60
+ "eval_samples_per_second": 339.518,
61
+ "eval_sokak_f1": 0.5461538461538461,
62
+ "eval_sokak_precision": 0.47651006711409394,
63
+ "eval_sokak_recall": 0.6396396396396397,
64
+ "eval_sokak_support": 111,
65
+ "eval_steps_per_second": 10.61,
66
+ "step": 48
67
+ },
68
+ {
69
+ "epoch": 2.0,
70
+ "learning_rate": 3.3333333333333335e-05,
71
+ "loss": 0.2457,
72
+ "step": 96
73
+ },
74
+ {
75
+ "epoch": 2.0,
76
+ "eval_Apartman/Site_f1": 0.6298342541436464,
77
+ "eval_Apartman/Site_precision": 0.5377358490566038,
78
+ "eval_Apartman/Site_recall": 0.76,
79
+ "eval_Apartman/Site_support": 75,
80
+ "eval_ad-soyad_f1": 0.7876447876447876,
81
+ "eval_ad-soyad_precision": 0.7445255474452555,
82
+ "eval_ad-soyad_recall": 0.8360655737704918,
83
+ "eval_ad-soyad_support": 122,
84
+ "eval_dis kapi no_f1": 0.0,
85
+ "eval_dis kapi no_precision": 0.0,
86
+ "eval_dis kapi no_recall": 0.0,
87
+ "eval_dis kapi no_support": 9,
88
+ "eval_ic kapi no_f1": 0.0,
89
+ "eval_ic kapi no_precision": 0.0,
90
+ "eval_ic kapi no_recall": 0.0,
91
+ "eval_ic kapi no_support": 10,
92
+ "eval_il_f1": 0.8622222222222222,
93
+ "eval_il_precision": 0.8151260504201681,
94
+ "eval_il_recall": 0.9150943396226415,
95
+ "eval_il_support": 106,
96
+ "eval_ilce_f1": 0.819672131147541,
97
+ "eval_ilce_precision": 0.7575757575757576,
98
+ "eval_ilce_recall": 0.8928571428571429,
99
+ "eval_ilce_support": 84,
100
+ "eval_kat_f1": 0.3157894736842105,
101
+ "eval_kat_precision": 0.24,
102
+ "eval_kat_recall": 0.46153846153846156,
103
+ "eval_kat_support": 13,
104
+ "eval_loss": 0.28523704409599304,
105
+ "eval_mahalle_f1": 0.6302521008403361,
106
+ "eval_mahalle_precision": 0.5952380952380952,
107
+ "eval_mahalle_recall": 0.6696428571428571,
108
+ "eval_mahalle_support": 112,
109
+ "eval_no_f1": 0.4509803921568628,
110
+ "eval_no_precision": 0.38333333333333336,
111
+ "eval_no_recall": 0.5476190476190477,
112
+ "eval_no_support": 42,
113
+ "eval_overall_accuracy": 0.9154565347125522,
114
+ "eval_overall_f1": 0.68733153638814,
115
+ "eval_overall_precision": 0.6375,
116
+ "eval_overall_recall": 0.7456140350877193,
117
+ "eval_runtime": 0.5594,
118
+ "eval_samples_per_second": 343.207,
119
+ "eval_sokak_f1": 0.6382978723404256,
120
+ "eval_sokak_precision": 0.6048387096774194,
121
+ "eval_sokak_recall": 0.6756756756756757,
122
+ "eval_sokak_support": 111,
123
+ "eval_steps_per_second": 10.725,
124
+ "step": 96
125
+ },
126
+ {
127
+ "epoch": 3.0,
128
+ "learning_rate": 2.2222222222222223e-05,
129
+ "loss": 0.1728,
130
+ "step": 144
131
+ },
132
+ {
133
+ "epoch": 3.0,
134
+ "eval_Apartman/Site_f1": 0.6436781609195402,
135
+ "eval_Apartman/Site_precision": 0.5656565656565656,
136
+ "eval_Apartman/Site_recall": 0.7466666666666667,
137
+ "eval_Apartman/Site_support": 75,
138
+ "eval_ad-soyad_f1": 0.7921568627450981,
139
+ "eval_ad-soyad_precision": 0.7593984962406015,
140
+ "eval_ad-soyad_recall": 0.8278688524590164,
141
+ "eval_ad-soyad_support": 122,
142
+ "eval_dis kapi no_f1": 0.0,
143
+ "eval_dis kapi no_precision": 0.0,
144
+ "eval_dis kapi no_recall": 0.0,
145
+ "eval_dis kapi no_support": 9,
146
+ "eval_ic kapi no_f1": 0.26666666666666666,
147
+ "eval_ic kapi no_precision": 0.4,
148
+ "eval_ic kapi no_recall": 0.2,
149
+ "eval_ic kapi no_support": 10,
150
+ "eval_il_f1": 0.9245283018867925,
151
+ "eval_il_precision": 0.9245283018867925,
152
+ "eval_il_recall": 0.9245283018867925,
153
+ "eval_il_support": 106,
154
+ "eval_ilce_f1": 0.8222222222222222,
155
+ "eval_ilce_precision": 0.7708333333333334,
156
+ "eval_ilce_recall": 0.8809523809523809,
157
+ "eval_ilce_support": 84,
158
+ "eval_kat_f1": 0.47058823529411764,
159
+ "eval_kat_precision": 0.38095238095238093,
160
+ "eval_kat_recall": 0.6153846153846154,
161
+ "eval_kat_support": 13,
162
+ "eval_loss": 0.25738969445228577,
163
+ "eval_mahalle_f1": 0.6297872340425532,
164
+ "eval_mahalle_precision": 0.6016260162601627,
165
+ "eval_mahalle_recall": 0.6607142857142857,
166
+ "eval_mahalle_support": 112,
167
+ "eval_no_f1": 0.37606837606837606,
168
+ "eval_no_precision": 0.29333333333333333,
169
+ "eval_no_recall": 0.5238095238095238,
170
+ "eval_no_support": 42,
171
+ "eval_overall_accuracy": 0.9242092699423116,
172
+ "eval_overall_f1": 0.7053206002728514,
173
+ "eval_overall_precision": 0.6611253196930946,
174
+ "eval_overall_recall": 0.7558479532163743,
175
+ "eval_runtime": 0.5539,
176
+ "eval_samples_per_second": 346.644,
177
+ "eval_sokak_f1": 0.6978723404255319,
178
+ "eval_sokak_precision": 0.6612903225806451,
179
+ "eval_sokak_recall": 0.7387387387387387,
180
+ "eval_sokak_support": 111,
181
+ "eval_steps_per_second": 10.833,
182
+ "step": 144
183
+ },
184
+ {
185
+ "epoch": 4.0,
186
+ "learning_rate": 1.1111111111111112e-05,
187
+ "loss": 0.1364,
188
+ "step": 192
189
+ },
190
+ {
191
+ "epoch": 4.0,
192
+ "eval_Apartman/Site_f1": 0.6540880503144655,
193
+ "eval_Apartman/Site_precision": 0.6190476190476191,
194
+ "eval_Apartman/Site_recall": 0.6933333333333334,
195
+ "eval_Apartman/Site_support": 75,
196
+ "eval_ad-soyad_f1": 0.8093385214007783,
197
+ "eval_ad-soyad_precision": 0.7703703703703704,
198
+ "eval_ad-soyad_recall": 0.8524590163934426,
199
+ "eval_ad-soyad_support": 122,
200
+ "eval_dis kapi no_f1": 0.0,
201
+ "eval_dis kapi no_precision": 0.0,
202
+ "eval_dis kapi no_recall": 0.0,
203
+ "eval_dis kapi no_support": 9,
204
+ "eval_ic kapi no_f1": 0.1818181818181818,
205
+ "eval_ic kapi no_precision": 0.16666666666666666,
206
+ "eval_ic kapi no_recall": 0.2,
207
+ "eval_ic kapi no_support": 10,
208
+ "eval_il_f1": 0.9194312796208531,
209
+ "eval_il_precision": 0.9238095238095239,
210
+ "eval_il_recall": 0.9150943396226415,
211
+ "eval_il_support": 106,
212
+ "eval_ilce_f1": 0.8287292817679559,
213
+ "eval_ilce_precision": 0.7731958762886598,
214
+ "eval_ilce_recall": 0.8928571428571429,
215
+ "eval_ilce_support": 84,
216
+ "eval_kat_f1": 0.4242424242424242,
217
+ "eval_kat_precision": 0.35,
218
+ "eval_kat_recall": 0.5384615384615384,
219
+ "eval_kat_support": 13,
220
+ "eval_loss": 0.252394437789917,
221
+ "eval_mahalle_f1": 0.6666666666666667,
222
+ "eval_mahalle_precision": 0.6470588235294118,
223
+ "eval_mahalle_recall": 0.6875,
224
+ "eval_mahalle_support": 112,
225
+ "eval_no_f1": 0.4571428571428571,
226
+ "eval_no_precision": 0.38095238095238093,
227
+ "eval_no_recall": 0.5714285714285714,
228
+ "eval_no_support": 42,
229
+ "eval_overall_accuracy": 0.926795305351104,
230
+ "eval_overall_f1": 0.708825573314802,
231
+ "eval_overall_precision": 0.6754966887417219,
232
+ "eval_overall_recall": 0.7456140350877193,
233
+ "eval_runtime": 0.5605,
234
+ "eval_samples_per_second": 342.58,
235
+ "eval_sokak_f1": 0.6233766233766234,
236
+ "eval_sokak_precision": 0.6,
237
+ "eval_sokak_recall": 0.6486486486486487,
238
+ "eval_sokak_support": 111,
239
+ "eval_steps_per_second": 10.706,
240
+ "step": 192
241
+ },
242
+ {
243
+ "epoch": 5.0,
244
+ "learning_rate": 0.0,
245
+ "loss": 0.1109,
246
+ "step": 240
247
+ },
248
+ {
249
+ "epoch": 5.0,
250
+ "eval_Apartman/Site_f1": 0.6538461538461539,
251
+ "eval_Apartman/Site_precision": 0.6296296296296297,
252
+ "eval_Apartman/Site_recall": 0.68,
253
+ "eval_Apartman/Site_support": 75,
254
+ "eval_ad-soyad_f1": 0.8142292490118578,
255
+ "eval_ad-soyad_precision": 0.7862595419847328,
256
+ "eval_ad-soyad_recall": 0.8442622950819673,
257
+ "eval_ad-soyad_support": 122,
258
+ "eval_dis kapi no_f1": 0.0,
259
+ "eval_dis kapi no_precision": 0.0,
260
+ "eval_dis kapi no_recall": 0.0,
261
+ "eval_dis kapi no_support": 9,
262
+ "eval_ic kapi no_f1": 0.2727272727272727,
263
+ "eval_ic kapi no_precision": 0.25,
264
+ "eval_ic kapi no_recall": 0.3,
265
+ "eval_ic kapi no_support": 10,
266
+ "eval_il_f1": 0.9142857142857143,
267
+ "eval_il_precision": 0.9230769230769231,
268
+ "eval_il_recall": 0.9056603773584906,
269
+ "eval_il_support": 106,
270
+ "eval_ilce_f1": 0.8379888268156424,
271
+ "eval_ilce_precision": 0.7894736842105263,
272
+ "eval_ilce_recall": 0.8928571428571429,
273
+ "eval_ilce_support": 84,
274
+ "eval_kat_f1": 0.4571428571428572,
275
+ "eval_kat_precision": 0.36363636363636365,
276
+ "eval_kat_recall": 0.6153846153846154,
277
+ "eval_kat_support": 13,
278
+ "eval_loss": 0.26428258419036865,
279
+ "eval_mahalle_f1": 0.6324786324786325,
280
+ "eval_mahalle_precision": 0.6065573770491803,
281
+ "eval_mahalle_recall": 0.6607142857142857,
282
+ "eval_mahalle_support": 112,
283
+ "eval_no_f1": 0.4752475247524753,
284
+ "eval_no_precision": 0.4067796610169492,
285
+ "eval_no_recall": 0.5714285714285714,
286
+ "eval_no_support": 42,
287
+ "eval_overall_accuracy": 0.9273920827531331,
288
+ "eval_overall_f1": 0.7178770949720671,
289
+ "eval_overall_precision": 0.6871657754010695,
290
+ "eval_overall_recall": 0.7514619883040936,
291
+ "eval_runtime": 0.5624,
292
+ "eval_samples_per_second": 341.388,
293
+ "eval_sokak_f1": 0.689655172413793,
294
+ "eval_sokak_precision": 0.6611570247933884,
295
+ "eval_sokak_recall": 0.7207207207207207,
296
+ "eval_sokak_support": 111,
297
+ "eval_steps_per_second": 10.668,
298
+ "step": 240
299
+ }
300
+ ],
301
+ "max_steps": 240,
302
+ "num_train_epochs": 5,
303
+ "total_flos": 178119641824452.0,
304
+ "trial_name": null,
305
+ "trial_params": null
306
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f0d70c5e94a38a22af269ef74c4de89c90b726eaac03d9491f3f70e5d12bd95a
3
+ size 3503
vocab.txt ADDED
The diff for this file is too large to render. See raw diff