Hugging Face's logo

stefan-it
/
hmbench-icdar-fr-hmbert-bs8-wsFalse-e10-lr3e-05-poolingfirst-layers-1-crfFalse-1

Token Classification
Flair
PyTorch
French
sequence-tagger-model
Model card Files Community
1
stefan-it commited on
Commit
62cc8c2
·
1 Parent(s): 6396c2f

Upload folder using huggingface_hub

Browse files
Files changed (5) hide show
  1. best-model.pt +3 -0
  2. dev.tsv +0 -0
  3. loss.tsv +11 -0
  4. test.tsv +0 -0
  5. training.log +241 -0
best-model.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9e0d7c22c0bbd01dc7de1887bd7101d9020ae3ffbe5b4160084ea75aaf31c92d
3
+ size 443311111
dev.tsv ADDED
The diff for this file is too large to render. See raw diff
 
loss.tsv ADDED
@@ -0,0 +1,11 @@
 
 
 
 
 
 
 
 
 
 
 
 
1
+ EPOCH TIMESTAMP LEARNING_RATE TRAIN_LOSS DEV_LOSS DEV_PRECISION DEV_RECALL DEV_F1 DEV_ACCURACY
2
+ 1 21:27:21 0.0000 0.4167 0.0992 0.6918 0.7262 0.7086 0.5702
3
+ 2 21:28:25 0.0000 0.1013 0.0805 0.7055 0.7749 0.7385 0.6004
4
+ 3 21:29:28 0.0000 0.0690 0.1053 0.7216 0.7828 0.7509 0.6234
5
+ 4 21:30:31 0.0000 0.0497 0.1206 0.7164 0.7771 0.7455 0.6134
6
+ 5 21:31:35 0.0000 0.0381 0.1374 0.7411 0.7771 0.7587 0.6343
7
+ 6 21:32:39 0.0000 0.0295 0.1686 0.7387 0.7771 0.7574 0.6274
8
+ 7 21:33:42 0.0000 0.0218 0.1858 0.7292 0.7919 0.7592 0.6318
9
+ 8 21:34:46 0.0000 0.0161 0.2050 0.7494 0.7681 0.7587 0.6293
10
+ 9 21:35:49 0.0000 0.0134 0.2013 0.7347 0.7862 0.7596 0.6330
11
+ 10 21:36:52 0.0000 0.0092 0.2173 0.7376 0.7726 0.7547 0.6255
test.tsv ADDED
The diff for this file is too large to render. See raw diff
 
training.log ADDED
@@ -0,0 +1,241 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ 2023-10-13 21:26:19,181 ----------------------------------------------------------------------------------------------------
2
+ 2023-10-13 21:26:19,182 Model: "SequenceTagger(
3
+ (embeddings): TransformerWordEmbeddings(
4
+ (model): BertModel(
5
+ (embeddings): BertEmbeddings(
6
+ (word_embeddings): Embedding(32001, 768)
7
+ (position_embeddings): Embedding(512, 768)
8
+ (token_type_embeddings): Embedding(2, 768)
9
+ (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
10
+ (dropout): Dropout(p=0.1, inplace=False)
11
+ )
12
+ (encoder): BertEncoder(
13
+ (layer): ModuleList(
14
+ (0-11): 12 x BertLayer(
15
+ (attention): BertAttention(
16
+ (self): BertSelfAttention(
17
+ (query): Linear(in_features=768, out_features=768, bias=True)
18
+ (key): Linear(in_features=768, out_features=768, bias=True)
19
+ (value): Linear(in_features=768, out_features=768, bias=True)
20
+ (dropout): Dropout(p=0.1, inplace=False)
21
+ )
22
+ (output): BertSelfOutput(
23
+ (dense): Linear(in_features=768, out_features=768, bias=True)
24
+ (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
25
+ (dropout): Dropout(p=0.1, inplace=False)
26
+ )
27
+ )
28
+ (intermediate): BertIntermediate(
29
+ (dense): Linear(in_features=768, out_features=3072, bias=True)
30
+ (intermediate_act_fn): GELUActivation()
31
+ )
32
+ (output): BertOutput(
33
+ (dense): Linear(in_features=3072, out_features=768, bias=True)
34
+ (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
35
+ (dropout): Dropout(p=0.1, inplace=False)
36
+ )
37
+ )
38
+ )
39
+ )
40
+ (pooler): BertPooler(
41
+ (dense): Linear(in_features=768, out_features=768, bias=True)
42
+ (activation): Tanh()
43
+ )
44
+ )
45
+ )
46
+ (locked_dropout): LockedDropout(p=0.5)
47
+ (linear): Linear(in_features=768, out_features=13, bias=True)
48
+ (loss_function): CrossEntropyLoss()
49
+ )"
50
+ 2023-10-13 21:26:19,182 ----------------------------------------------------------------------------------------------------
51
+ 2023-10-13 21:26:19,182 MultiCorpus: 7936 train + 992 dev + 992 test sentences
52
+ - NER_ICDAR_EUROPEANA Corpus: 7936 train + 992 dev + 992 test sentences - /root/.flair/datasets/ner_icdar_europeana/fr
53
+ 2023-10-13 21:26:19,182 ----------------------------------------------------------------------------------------------------
54
+ 2023-10-13 21:26:19,182 Train: 7936 sentences
55
+ 2023-10-13 21:26:19,182 (train_with_dev=False, train_with_test=False)
56
+ 2023-10-13 21:26:19,182 ----------------------------------------------------------------------------------------------------
57
+ 2023-10-13 21:26:19,182 Training Params:
58
+ 2023-10-13 21:26:19,182 - learning_rate: "3e-05"
59
+ 2023-10-13 21:26:19,182 - mini_batch_size: "8"
60
+ 2023-10-13 21:26:19,182 - max_epochs: "10"
61
+ 2023-10-13 21:26:19,182 - shuffle: "True"
62
+ 2023-10-13 21:26:19,182 ----------------------------------------------------------------------------------------------------
63
+ 2023-10-13 21:26:19,182 Plugins:
64
+ 2023-10-13 21:26:19,182 - LinearScheduler | warmup_fraction: '0.1'
65
+ 2023-10-13 21:26:19,182 ----------------------------------------------------------------------------------------------------
66
+ 2023-10-13 21:26:19,182 Final evaluation on model from best epoch (best-model.pt)
67
+ 2023-10-13 21:26:19,182 - metric: "('micro avg', 'f1-score')"
68
+ 2023-10-13 21:26:19,182 ----------------------------------------------------------------------------------------------------
69
+ 2023-10-13 21:26:19,182 Computation:
70
+ 2023-10-13 21:26:19,182 - compute on device: cuda:0
71
+ 2023-10-13 21:26:19,182 - embedding storage: none
72
+ 2023-10-13 21:26:19,182 ----------------------------------------------------------------------------------------------------
73
+ 2023-10-13 21:26:19,183 Model training base path: "hmbench-icdar/fr-dbmdz/bert-base-historic-multilingual-cased-bs8-wsFalse-e10-lr3e-05-poolingfirst-layers-1-crfFalse-1"
74
+ 2023-10-13 21:26:19,183 ----------------------------------------------------------------------------------------------------
75
+ 2023-10-13 21:26:19,183 ----------------------------------------------------------------------------------------------------
76
+ 2023-10-13 21:26:25,065 epoch 1 - iter 99/992 - loss 2.28557417 - time (sec): 5.88 - samples/sec: 2733.74 - lr: 0.000003 - momentum: 0.000000
77
+ 2023-10-13 21:26:31,062 epoch 1 - iter 198/992 - loss 1.39151134 - time (sec): 11.88 - samples/sec: 2741.70 - lr: 0.000006 - momentum: 0.000000
78
+ 2023-10-13 21:26:36,969 epoch 1 - iter 297/992 - loss 1.02546157 - time (sec): 17.79 - samples/sec: 2764.99 - lr: 0.000009 - momentum: 0.000000
79
+ 2023-10-13 21:26:42,777 epoch 1 - iter 396/992 - loss 0.82409225 - time (sec): 23.59 - samples/sec: 2767.68 - lr: 0.000012 - momentum: 0.000000
80
+ 2023-10-13 21:26:48,467 epoch 1 - iter 495/992 - loss 0.69589370 - time (sec): 29.28 - samples/sec: 2780.04 - lr: 0.000015 - momentum: 0.000000
81
+ 2023-10-13 21:26:54,198 epoch 1 - iter 594/992 - loss 0.60583632 - time (sec): 35.01 - samples/sec: 2786.20 - lr: 0.000018 - momentum: 0.000000
82
+ 2023-10-13 21:27:00,081 epoch 1 - iter 693/992 - loss 0.53770132 - time (sec): 40.90 - samples/sec: 2804.16 - lr: 0.000021 - momentum: 0.000000
83
+ 2023-10-13 21:27:06,207 epoch 1 - iter 792/992 - loss 0.48422796 - time (sec): 47.02 - samples/sec: 2810.64 - lr: 0.000024 - momentum: 0.000000
84
+ 2023-10-13 21:27:12,012 epoch 1 - iter 891/992 - loss 0.44737479 - time (sec): 52.83 - samples/sec: 2803.31 - lr: 0.000027 - momentum: 0.000000
85
+ 2023-10-13 21:27:18,204 epoch 1 - iter 990/992 - loss 0.41710425 - time (sec): 59.02 - samples/sec: 2774.99 - lr: 0.000030 - momentum: 0.000000
86
+ 2023-10-13 21:27:18,322 ----------------------------------------------------------------------------------------------------
87
+ 2023-10-13 21:27:18,322 EPOCH 1 done: loss 0.4167 - lr: 0.000030
88
+ 2023-10-13 21:27:21,503 DEV : loss 0.09923805296421051 - f1-score (micro avg) 0.7086
89
+ 2023-10-13 21:27:21,525 saving best model
90
+ 2023-10-13 21:27:21,981 ----------------------------------------------------------------------------------------------------
91
+ 2023-10-13 21:27:28,019 epoch 2 - iter 99/992 - loss 0.11613135 - time (sec): 6.04 - samples/sec: 2859.34 - lr: 0.000030 - momentum: 0.000000
92
+ 2023-10-13 21:27:33,706 epoch 2 - iter 198/992 - loss 0.11269936 - time (sec): 11.72 - samples/sec: 2765.83 - lr: 0.000029 - momentum: 0.000000
93
+ 2023-10-13 21:27:40,128 epoch 2 - iter 297/992 - loss 0.10852299 - time (sec): 18.14 - samples/sec: 2746.57 - lr: 0.000029 - momentum: 0.000000
94
+ 2023-10-13 21:27:45,920 epoch 2 - iter 396/992 - loss 0.10649798 - time (sec): 23.94 - samples/sec: 2705.05 - lr: 0.000029 - momentum: 0.000000
95
+ 2023-10-13 21:27:51,989 epoch 2 - iter 495/992 - loss 0.10652407 - time (sec): 30.01 - samples/sec: 2737.62 - lr: 0.000028 - momentum: 0.000000
96
+ 2023-10-13 21:27:57,928 epoch 2 - iter 594/992 - loss 0.10483546 - time (sec): 35.94 - samples/sec: 2729.34 - lr: 0.000028 - momentum: 0.000000
97
+ 2023-10-13 21:28:04,046 epoch 2 - iter 693/992 - loss 0.10404519 - time (sec): 42.06 - samples/sec: 2724.36 - lr: 0.000028 - momentum: 0.000000
98
+ 2023-10-13 21:28:09,845 epoch 2 - iter 792/992 - loss 0.10351903 - time (sec): 47.86 - samples/sec: 2727.54 - lr: 0.000027 - momentum: 0.000000
99
+ 2023-10-13 21:28:15,662 epoch 2 - iter 891/992 - loss 0.10258918 - time (sec): 53.68 - samples/sec: 2744.25 - lr: 0.000027 - momentum: 0.000000
100
+ 2023-10-13 21:28:21,494 epoch 2 - iter 990/992 - loss 0.10133427 - time (sec): 59.51 - samples/sec: 2751.72 - lr: 0.000027 - momentum: 0.000000
101
+ 2023-10-13 21:28:21,607 ----------------------------------------------------------------------------------------------------
102
+ 2023-10-13 21:28:21,607 EPOCH 2 done: loss 0.1013 - lr: 0.000027
103
+ 2023-10-13 21:28:25,515 DEV : loss 0.0804639682173729 - f1-score (micro avg) 0.7385
104
+ 2023-10-13 21:28:25,535 saving best model
105
+ 2023-10-13 21:28:26,011 ----------------------------------------------------------------------------------------------------
106
+ 2023-10-13 21:28:31,880 epoch 3 - iter 99/992 - loss 0.06775603 - time (sec): 5.87 - samples/sec: 2791.67 - lr: 0.000026 - momentum: 0.000000
107
+ 2023-10-13 21:28:37,828 epoch 3 - iter 198/992 - loss 0.06948443 - time (sec): 11.82 - samples/sec: 2698.32 - lr: 0.000026 - momentum: 0.000000
108
+ 2023-10-13 21:28:43,969 epoch 3 - iter 297/992 - loss 0.06447600 - time (sec): 17.96 - samples/sec: 2753.14 - lr: 0.000026 - momentum: 0.000000
109
+ 2023-10-13 21:28:50,058 epoch 3 - iter 396/992 - loss 0.06751946 - time (sec): 24.05 - samples/sec: 2771.64 - lr: 0.000025 - momentum: 0.000000
110
+ 2023-10-13 21:28:55,993 epoch 3 - iter 495/992 - loss 0.06685992 - time (sec): 29.98 - samples/sec: 2763.41 - lr: 0.000025 - momentum: 0.000000
111
+ 2023-10-13 21:29:01,692 epoch 3 - iter 594/992 - loss 0.06777144 - time (sec): 35.68 - samples/sec: 2775.56 - lr: 0.000025 - momentum: 0.000000
112
+ 2023-10-13 21:29:07,475 epoch 3 - iter 693/992 - loss 0.06708595 - time (sec): 41.46 - samples/sec: 2773.37 - lr: 0.000024 - momentum: 0.000000
113
+ 2023-10-13 21:29:13,244 epoch 3 - iter 792/992 - loss 0.06891828 - time (sec): 47.23 - samples/sec: 2784.55 - lr: 0.000024 - momentum: 0.000000
114
+ 2023-10-13 21:29:19,038 epoch 3 - iter 891/992 - loss 0.06968673 - time (sec): 53.03 - samples/sec: 2786.41 - lr: 0.000024 - momentum: 0.000000
115
+ 2023-10-13 21:29:24,682 epoch 3 - iter 990/992 - loss 0.06906410 - time (sec): 58.67 - samples/sec: 2787.68 - lr: 0.000023 - momentum: 0.000000
116
+ 2023-10-13 21:29:24,800 ----------------------------------------------------------------------------------------------------
117
+ 2023-10-13 21:29:24,801 EPOCH 3 done: loss 0.0690 - lr: 0.000023
118
+ 2023-10-13 21:29:28,299 DEV : loss 0.1052832305431366 - f1-score (micro avg) 0.7509
119
+ 2023-10-13 21:29:28,329 saving best model
120
+ 2023-10-13 21:29:28,851 ----------------------------------------------------------------------------------------------------
121
+ 2023-10-13 21:29:35,125 epoch 4 - iter 99/992 - loss 0.04809586 - time (sec): 6.27 - samples/sec: 2636.57 - lr: 0.000023 - momentum: 0.000000
122
+ 2023-10-13 21:29:41,100 epoch 4 - iter 198/992 - loss 0.04947603 - time (sec): 12.25 - samples/sec: 2714.99 - lr: 0.000023 - momentum: 0.000000
123
+ 2023-10-13 21:29:46,795 epoch 4 - iter 297/992 - loss 0.05118762 - time (sec): 17.94 - samples/sec: 2734.62 - lr: 0.000022 - momentum: 0.000000
124
+ 2023-10-13 21:29:52,796 epoch 4 - iter 396/992 - loss 0.04965416 - time (sec): 23.94 - samples/sec: 2717.92 - lr: 0.000022 - momentum: 0.000000
125
+ 2023-10-13 21:29:58,771 epoch 4 - iter 495/992 - loss 0.04848418 - time (sec): 29.92 - samples/sec: 2721.76 - lr: 0.000022 - momentum: 0.000000
126
+ 2023-10-13 21:30:04,653 epoch 4 - iter 594/992 - loss 0.04858443 - time (sec): 35.80 - samples/sec: 2729.07 - lr: 0.000021 - momentum: 0.000000
127
+ 2023-10-13 21:30:10,562 epoch 4 - iter 693/992 - loss 0.04967146 - time (sec): 41.71 - samples/sec: 2732.89 - lr: 0.000021 - momentum: 0.000000
128
+ 2023-10-13 21:30:16,426 epoch 4 - iter 792/992 - loss 0.04938520 - time (sec): 47.57 - samples/sec: 2731.92 - lr: 0.000021 - momentum: 0.000000
129
+ 2023-10-13 21:30:22,330 epoch 4 - iter 891/992 - loss 0.04876000 - time (sec): 53.48 - samples/sec: 2740.55 - lr: 0.000020 - momentum: 0.000000
130
+ 2023-10-13 21:30:28,272 epoch 4 - iter 990/992 - loss 0.04971305 - time (sec): 59.42 - samples/sec: 2755.83 - lr: 0.000020 - momentum: 0.000000
131
+ 2023-10-13 21:30:28,383 ----------------------------------------------------------------------------------------------------
132
+ 2023-10-13 21:30:28,383 EPOCH 4 done: loss 0.0497 - lr: 0.000020
133
+ 2023-10-13 21:30:31,897 DEV : loss 0.12055560946464539 - f1-score (micro avg) 0.7455
134
+ 2023-10-13 21:30:31,919 ----------------------------------------------------------------------------------------------------
135
+ 2023-10-13 21:30:37,673 epoch 5 - iter 99/992 - loss 0.04227405 - time (sec): 5.75 - samples/sec: 2891.99 - lr: 0.000020 - momentum: 0.000000
136
+ 2023-10-13 21:30:43,513 epoch 5 - iter 198/992 - loss 0.03608106 - time (sec): 11.59 - samples/sec: 2842.42 - lr: 0.000019 - momentum: 0.000000
137
+ 2023-10-13 21:30:49,376 epoch 5 - iter 297/992 - loss 0.03886863 - time (sec): 17.46 - samples/sec: 2822.33 - lr: 0.000019 - momentum: 0.000000
138
+ 2023-10-13 21:30:56,245 epoch 5 - iter 396/992 - loss 0.03816738 - time (sec): 24.32 - samples/sec: 2740.89 - lr: 0.000019 - momentum: 0.000000
139
+ 2023-10-13 21:31:02,116 epoch 5 - iter 495/992 - loss 0.03804566 - time (sec): 30.20 - samples/sec: 2755.43 - lr: 0.000018 - momentum: 0.000000
140
+ 2023-10-13 21:31:07,852 epoch 5 - iter 594/992 - loss 0.03689071 - time (sec): 35.93 - samples/sec: 2761.47 - lr: 0.000018 - momentum: 0.000000
141
+ 2023-10-13 21:31:13,786 epoch 5 - iter 693/992 - loss 0.03872646 - time (sec): 41.87 - samples/sec: 2749.53 - lr: 0.000018 - momentum: 0.000000
142
+ 2023-10-13 21:31:19,687 epoch 5 - iter 792/992 - loss 0.03754210 - time (sec): 47.77 - samples/sec: 2754.69 - lr: 0.000017 - momentum: 0.000000
143
+ 2023-10-13 21:31:25,930 epoch 5 - iter 891/992 - loss 0.03755859 - time (sec): 54.01 - samples/sec: 2741.27 - lr: 0.000017 - momentum: 0.000000
144
+ 2023-10-13 21:31:31,815 epoch 5 - iter 990/992 - loss 0.03809129 - time (sec): 59.89 - samples/sec: 2731.36 - lr: 0.000017 - momentum: 0.000000
145
+ 2023-10-13 21:31:31,954 ----------------------------------------------------------------------------------------------------
146
+ 2023-10-13 21:31:31,954 EPOCH 5 done: loss 0.0381 - lr: 0.000017
147
+ 2023-10-13 21:31:35,520 DEV : loss 0.13742151856422424 - f1-score (micro avg) 0.7587
148
+ 2023-10-13 21:31:35,544 saving best model
149
+ 2023-10-13 21:31:36,125 ----------------------------------------------------------------------------------------------------
150
+ 2023-10-13 21:31:42,337 epoch 6 - iter 99/992 - loss 0.02712909 - time (sec): 6.21 - samples/sec: 2670.73 - lr: 0.000016 - momentum: 0.000000
151
+ 2023-10-13 21:31:48,728 epoch 6 - iter 198/992 - loss 0.02785322 - time (sec): 12.60 - samples/sec: 2683.50 - lr: 0.000016 - momentum: 0.000000
152
+ 2023-10-13 21:31:54,715 epoch 6 - iter 297/992 - loss 0.02848273 - time (sec): 18.59 - samples/sec: 2672.94 - lr: 0.000016 - momentum: 0.000000
153
+ 2023-10-13 21:32:00,706 epoch 6 - iter 396/992 - loss 0.02785739 - time (sec): 24.58 - samples/sec: 2667.29 - lr: 0.000015 - momentum: 0.000000
154
+ 2023-10-13 21:32:06,591 epoch 6 - iter 495/992 - loss 0.02751215 - time (sec): 30.46 - samples/sec: 2689.46 - lr: 0.000015 - momentum: 0.000000
155
+ 2023-10-13 21:32:12,351 epoch 6 - iter 594/992 - loss 0.02798228 - time (sec): 36.22 - samples/sec: 2715.72 - lr: 0.000015 - momentum: 0.000000
156
+ 2023-10-13 21:32:18,302 epoch 6 - iter 693/992 - loss 0.02829397 - time (sec): 42.18 - samples/sec: 2730.18 - lr: 0.000014 - momentum: 0.000000
157
+ 2023-10-13 21:32:24,156 epoch 6 - iter 792/992 - loss 0.02864477 - time (sec): 48.03 - samples/sec: 2736.41 - lr: 0.000014 - momentum: 0.000000
158
+ 2023-10-13 21:32:30,000 epoch 6 - iter 891/992 - loss 0.02875296 - time (sec): 53.87 - samples/sec: 2736.42 - lr: 0.000014 - momentum: 0.000000
159
+ 2023-10-13 21:32:36,022 epoch 6 - iter 990/992 - loss 0.02942813 - time (sec): 59.90 - samples/sec: 2732.62 - lr: 0.000013 - momentum: 0.000000
160
+ 2023-10-13 21:32:36,132 ----------------------------------------------------------------------------------------------------
161
+ 2023-10-13 21:32:36,132 EPOCH 6 done: loss 0.0295 - lr: 0.000013
162
+ 2023-10-13 21:32:39,605 DEV : loss 0.16862468421459198 - f1-score (micro avg) 0.7574
163
+ 2023-10-13 21:32:39,626 ----------------------------------------------------------------------------------------------------
164
+ 2023-10-13 21:32:45,397 epoch 7 - iter 99/992 - loss 0.02084181 - time (sec): 5.77 - samples/sec: 2663.18 - lr: 0.000013 - momentum: 0.000000
165
+ 2023-10-13 21:32:51,692 epoch 7 - iter 198/992 - loss 0.02364712 - time (sec): 12.07 - samples/sec: 2669.23 - lr: 0.000013 - momentum: 0.000000
166
+ 2023-10-13 21:32:57,470 epoch 7 - iter 297/992 - loss 0.02055440 - time (sec): 17.84 - samples/sec: 2677.99 - lr: 0.000012 - momentum: 0.000000
167
+ 2023-10-13 21:33:03,230 epoch 7 - iter 396/992 - loss 0.02123931 - time (sec): 23.60 - samples/sec: 2709.80 - lr: 0.000012 - momentum: 0.000000
168
+ 2023-10-13 21:33:09,233 epoch 7 - iter 495/992 - loss 0.02161191 - time (sec): 29.61 - samples/sec: 2740.96 - lr: 0.000012 - momentum: 0.000000
169
+ 2023-10-13 21:33:15,904 epoch 7 - iter 594/992 - loss 0.02114814 - time (sec): 36.28 - samples/sec: 2714.40 - lr: 0.000011 - momentum: 0.000000
170
+ 2023-10-13 21:33:21,533 epoch 7 - iter 693/992 - loss 0.02119562 - time (sec): 41.91 - samples/sec: 2712.74 - lr: 0.000011 - momentum: 0.000000
171
+ 2023-10-13 21:33:27,488 epoch 7 - iter 792/992 - loss 0.02088100 - time (sec): 47.86 - samples/sec: 2730.44 - lr: 0.000011 - momentum: 0.000000
172
+ 2023-10-13 21:33:33,206 epoch 7 - iter 891/992 - loss 0.02187657 - time (sec): 53.58 - samples/sec: 2744.28 - lr: 0.000010 - momentum: 0.000000
173
+ 2023-10-13 21:33:38,997 epoch 7 - iter 990/992 - loss 0.02186821 - time (sec): 59.37 - samples/sec: 2755.54 - lr: 0.000010 - momentum: 0.000000
174
+ 2023-10-13 21:33:39,128 ----------------------------------------------------------------------------------------------------
175
+ 2023-10-13 21:33:39,129 EPOCH 7 done: loss 0.0218 - lr: 0.000010
176
+ 2023-10-13 21:33:42,635 DEV : loss 0.18575911223888397 - f1-score (micro avg) 0.7592
177
+ 2023-10-13 21:33:42,667 saving best model
178
+ 2023-10-13 21:33:43,180 ----------------------------------------------------------------------------------------------------
179
+ 2023-10-13 21:33:49,709 epoch 8 - iter 99/992 - loss 0.01242652 - time (sec): 6.52 - samples/sec: 2504.77 - lr: 0.000010 - momentum: 0.000000
180
+ 2023-10-13 21:33:55,885 epoch 8 - iter 198/992 - loss 0.01492552 - time (sec): 12.70 - samples/sec: 2595.31 - lr: 0.000009 - momentum: 0.000000
181
+ 2023-10-13 21:34:01,748 epoch 8 - iter 297/992 - loss 0.01415991 - time (sec): 18.56 - samples/sec: 2631.07 - lr: 0.000009 - momentum: 0.000000
182
+ 2023-10-13 21:34:07,558 epoch 8 - iter 396/992 - loss 0.01548353 - time (sec): 24.37 - samples/sec: 2700.18 - lr: 0.000009 - momentum: 0.000000
183
+ 2023-10-13 21:34:13,303 epoch 8 - iter 495/992 - loss 0.01575741 - time (sec): 30.12 - samples/sec: 2718.52 - lr: 0.000008 - momentum: 0.000000
184
+ 2023-10-13 21:34:18,907 epoch 8 - iter 594/992 - loss 0.01594169 - time (sec): 35.72 - samples/sec: 2723.58 - lr: 0.000008 - momentum: 0.000000
185
+ 2023-10-13 21:34:24,857 epoch 8 - iter 693/992 - loss 0.01654981 - time (sec): 41.67 - samples/sec: 2735.99 - lr: 0.000008 - momentum: 0.000000
186
+ 2023-10-13 21:34:31,096 epoch 8 - iter 792/992 - loss 0.01587979 - time (sec): 47.91 - samples/sec: 2746.56 - lr: 0.000007 - momentum: 0.000000
187
+ 2023-10-13 21:34:36,972 epoch 8 - iter 891/992 - loss 0.01575634 - time (sec): 53.79 - samples/sec: 2741.67 - lr: 0.000007 - momentum: 0.000000
188
+ 2023-10-13 21:34:42,829 epoch 8 - iter 990/992 - loss 0.01614502 - time (sec): 59.64 - samples/sec: 2745.24 - lr: 0.000007 - momentum: 0.000000
189
+ 2023-10-13 21:34:42,936 ----------------------------------------------------------------------------------------------------
190
+ 2023-10-13 21:34:42,936 EPOCH 8 done: loss 0.0161 - lr: 0.000007
191
+ 2023-10-13 21:34:46,404 DEV : loss 0.20498403906822205 - f1-score (micro avg) 0.7587
192
+ 2023-10-13 21:34:46,426 ----------------------------------------------------------------------------------------------------
193
+ 2023-10-13 21:34:52,358 epoch 9 - iter 99/992 - loss 0.01356369 - time (sec): 5.93 - samples/sec: 2823.61 - lr: 0.000006 - momentum: 0.000000
194
+ 2023-10-13 21:34:58,266 epoch 9 - iter 198/992 - loss 0.01060758 - time (sec): 11.84 - samples/sec: 2769.20 - lr: 0.000006 - momentum: 0.000000
195
+ 2023-10-13 21:35:03,917 epoch 9 - iter 297/992 - loss 0.01135131 - time (sec): 17.49 - samples/sec: 2781.11 - lr: 0.000006 - momentum: 0.000000
196
+ 2023-10-13 21:35:09,966 epoch 9 - iter 396/992 - loss 0.01091601 - time (sec): 23.54 - samples/sec: 2794.35 - lr: 0.000005 - momentum: 0.000000
197
+ 2023-10-13 21:35:15,776 epoch 9 - iter 495/992 - loss 0.01147457 - time (sec): 29.35 - samples/sec: 2811.81 - lr: 0.000005 - momentum: 0.000000
198
+ 2023-10-13 21:35:21,554 epoch 9 - iter 594/992 - loss 0.01149657 - time (sec): 35.13 - samples/sec: 2792.52 - lr: 0.000005 - momentum: 0.000000
199
+ 2023-10-13 21:35:27,248 epoch 9 - iter 693/992 - loss 0.01184055 - time (sec): 40.82 - samples/sec: 2796.65 - lr: 0.000004 - momentum: 0.000000
200
+ 2023-10-13 21:35:33,220 epoch 9 - iter 792/992 - loss 0.01192689 - time (sec): 46.79 - samples/sec: 2799.57 - lr: 0.000004 - momentum: 0.000000
201
+ 2023-10-13 21:35:39,316 epoch 9 - iter 891/992 - loss 0.01323888 - time (sec): 52.89 - samples/sec: 2795.24 - lr: 0.000004 - momentum: 0.000000
202
+ 2023-10-13 21:35:45,243 epoch 9 - iter 990/992 - loss 0.01337615 - time (sec): 58.82 - samples/sec: 2781.18 - lr: 0.000003 - momentum: 0.000000
203
+ 2023-10-13 21:35:45,385 ----------------------------------------------------------------------------------------------------
204
+ 2023-10-13 21:35:45,385 EPOCH 9 done: loss 0.0134 - lr: 0.000003
205
+ 2023-10-13 21:35:49,385 DEV : loss 0.20130668580532074 - f1-score (micro avg) 0.7596
206
+ 2023-10-13 21:35:49,409 saving best model
207
+ 2023-10-13 21:35:49,998 ----------------------------------------------------------------------------------------------------
208
+ 2023-10-13 21:35:56,107 epoch 10 - iter 99/992 - loss 0.01231225 - time (sec): 6.11 - samples/sec: 2791.64 - lr: 0.000003 - momentum: 0.000000
209
+ 2023-10-13 21:36:01,882 epoch 10 - iter 198/992 - loss 0.00908713 - time (sec): 11.88 - samples/sec: 2758.24 - lr: 0.000003 - momentum: 0.000000
210
+ 2023-10-13 21:36:07,740 epoch 10 - iter 297/992 - loss 0.00865854 - time (sec): 17.74 - samples/sec: 2751.19 - lr: 0.000002 - momentum: 0.000000
211
+ 2023-10-13 21:36:13,958 epoch 10 - iter 396/992 - loss 0.00923764 - time (sec): 23.96 - samples/sec: 2731.85 - lr: 0.000002 - momentum: 0.000000
212
+ 2023-10-13 21:36:19,594 epoch 10 - iter 495/992 - loss 0.00934583 - time (sec): 29.59 - samples/sec: 2753.35 - lr: 0.000002 - momentum: 0.000000
213
+ 2023-10-13 21:36:25,379 epoch 10 - iter 594/992 - loss 0.00873248 - time (sec): 35.38 - samples/sec: 2769.19 - lr: 0.000001 - momentum: 0.000000
214
+ 2023-10-13 21:36:31,203 epoch 10 - iter 693/992 - loss 0.00930823 - time (sec): 41.20 - samples/sec: 2774.68 - lr: 0.000001 - momentum: 0.000000
215
+ 2023-10-13 21:36:37,052 epoch 10 - iter 792/992 - loss 0.00939506 - time (sec): 47.05 - samples/sec: 2782.82 - lr: 0.000001 - momentum: 0.000000
216
+ 2023-10-13 21:36:42,967 epoch 10 - iter 891/992 - loss 0.00942042 - time (sec): 52.97 - samples/sec: 2794.26 - lr: 0.000000 - momentum: 0.000000
217
+ 2023-10-13 21:36:48,640 epoch 10 - iter 990/992 - loss 0.00923242 - time (sec): 58.64 - samples/sec: 2790.71 - lr: 0.000000 - momentum: 0.000000
218
+ 2023-10-13 21:36:48,750 ----------------------------------------------------------------------------------------------------
219
+ 2023-10-13 21:36:48,750 EPOCH 10 done: loss 0.0092 - lr: 0.000000
220
+ 2023-10-13 21:36:52,276 DEV : loss 0.21729230880737305 - f1-score (micro avg) 0.7547
221
+ 2023-10-13 21:36:52,716 ----------------------------------------------------------------------------------------------------
222
+ 2023-10-13 21:36:52,717 Loading model from best epoch ...
223
+ 2023-10-13 21:36:54,254 SequenceTagger predicts: Dictionary with 13 tags: O, S-PER, B-PER, E-PER, I-PER, S-LOC, B-LOC, E-LOC, I-LOC, S-ORG, B-ORG, E-ORG, I-ORG
224
+ 2023-10-13 21:36:57,774
225
+ Results:
226
+ - F-score (micro) 0.7742
227
+ - F-score (macro) 0.7002
228
+ - Accuracy 0.6494
229
+
230
+ By class:
231
+ precision recall f1-score support
232
+
233
+ LOC 0.7980 0.8504 0.8234 655
234
+ PER 0.7312 0.8296 0.7773 223
235
+ ORG 0.5124 0.4882 0.5000 127
236
+
237
+ micro avg 0.7500 0.8000 0.7742 1005
238
+ macro avg 0.6805 0.7227 0.7002 1005
239
+ weighted avg 0.7471 0.8000 0.7723 1005
240
+
241
+ 2023-10-13 21:36:57,774 ----------------------------------------------------------------------------------------------------