Hugging Face's logo

stefan-it
/
hmbench-newseye-de-hmbert-bs4-wsFalse-e10-lr5e-05-poolingfirst-layers-1-crfFalse-3

Token Classification
Flair
PyTorch
German
sequence-tagger-model
Model card Files Community
2
stefan-it commited on
Commit
f386d9e
•
1 Parent(s): 5d50e08

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 +242 -0
best-model.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:196a32731dac432aae4f6962582b25a2795ac3206f2bd9cbb472d570e03c74b1
3
+ size 443323527
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 18:35:14 0.0000 0.3590 0.1936 0.1774 0.4602 0.2561 0.1482
3
+ 2 18:39:38 0.0000 0.1966 0.2665 0.2294 0.4905 0.3126 0.1871
4
+ 3 18:43:59 0.0000 0.1452 0.2110 0.2211 0.5644 0.3177 0.1897
5
+ 4 18:48:23 0.0000 0.1100 0.3511 0.2156 0.6117 0.3189 0.1908
6
+ 5 18:52:48 0.0000 0.0784 0.2648 0.2555 0.5682 0.3525 0.2155
7
+ 6 18:57:11 0.0000 0.0632 0.3284 0.2598 0.5265 0.3479 0.2127
8
+ 7 19:01:33 0.0000 0.0437 0.3804 0.2531 0.5720 0.3510 0.2139
9
+ 8 19:05:59 0.0000 0.0330 0.3673 0.2916 0.5644 0.3845 0.2395
10
+ 9 19:10:28 0.0000 0.0232 0.4655 0.2504 0.5966 0.3527 0.2160
11
+ 10 19:14:53 0.0000 0.0148 0.4226 0.2778 0.5682 0.3731 0.2311
test.tsv ADDED
The diff for this file is too large to render. See raw diff
 
training.log ADDED
@@ -0,0 +1,242 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ 2023-10-15 18:30:53,960 ----------------------------------------------------------------------------------------------------
2
+ 2023-10-15 18:30:53,961 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=17, bias=True)
48
+ (loss_function): CrossEntropyLoss()
49
+ )"
50
+ 2023-10-15 18:30:53,961 ----------------------------------------------------------------------------------------------------
51
+ 2023-10-15 18:30:53,961 MultiCorpus: 20847 train + 1123 dev + 3350 test sentences
52
+ - NER_HIPE_2022 Corpus: 20847 train + 1123 dev + 3350 test sentences - /root/.flair/datasets/ner_hipe_2022/v2.1/newseye/de/with_doc_seperator
53
+ 2023-10-15 18:30:53,961 ----------------------------------------------------------------------------------------------------
54
+ 2023-10-15 18:30:53,961 Train: 20847 sentences
55
+ 2023-10-15 18:30:53,961 (train_with_dev=False, train_with_test=False)
56
+ 2023-10-15 18:30:53,961 ----------------------------------------------------------------------------------------------------
57
+ 2023-10-15 18:30:53,961 Training Params:
58
+ 2023-10-15 18:30:53,961 - learning_rate: "5e-05"
59
+ 2023-10-15 18:30:53,961 - mini_batch_size: "4"
60
+ 2023-10-15 18:30:53,961 - max_epochs: "10"
61
+ 2023-10-15 18:30:53,961 - shuffle: "True"
62
+ 2023-10-15 18:30:53,961 ----------------------------------------------------------------------------------------------------
63
+ 2023-10-15 18:30:53,961 Plugins:
64
+ 2023-10-15 18:30:53,961 - LinearScheduler | warmup_fraction: '0.1'
65
+ 2023-10-15 18:30:53,961 ----------------------------------------------------------------------------------------------------
66
+ 2023-10-15 18:30:53,961 Final evaluation on model from best epoch (best-model.pt)
67
+ 2023-10-15 18:30:53,961 - metric: "('micro avg', 'f1-score')"
68
+ 2023-10-15 18:30:53,962 ----------------------------------------------------------------------------------------------------
69
+ 2023-10-15 18:30:53,962 Computation:
70
+ 2023-10-15 18:30:53,962 - compute on device: cuda:0
71
+ 2023-10-15 18:30:53,962 - embedding storage: none
72
+ 2023-10-15 18:30:53,962 ----------------------------------------------------------------------------------------------------
73
+ 2023-10-15 18:30:53,962 Model training base path: "hmbench-newseye/de-dbmdz/bert-base-historic-multilingual-cased-bs4-wsFalse-e10-lr5e-05-poolingfirst-layers-1-crfFalse-3"
74
+ 2023-10-15 18:30:53,962 ----------------------------------------------------------------------------------------------------
75
+ 2023-10-15 18:30:53,962 ----------------------------------------------------------------------------------------------------
76
+ 2023-10-15 18:31:18,728 epoch 1 - iter 521/5212 - loss 1.28960364 - time (sec): 24.77 - samples/sec: 1414.43 - lr: 0.000005 - momentum: 0.000000
77
+ 2023-10-15 18:31:44,032 epoch 1 - iter 1042/5212 - loss 0.87638561 - time (sec): 50.07 - samples/sec: 1417.33 - lr: 0.000010 - momentum: 0.000000
78
+ 2023-10-15 18:32:10,016 epoch 1 - iter 1563/5212 - loss 0.66894164 - time (sec): 76.05 - samples/sec: 1423.01 - lr: 0.000015 - momentum: 0.000000
79
+ 2023-10-15 18:32:35,647 epoch 1 - iter 2084/5212 - loss 0.55810972 - time (sec): 101.68 - samples/sec: 1433.24 - lr: 0.000020 - momentum: 0.000000
80
+ 2023-10-15 18:33:00,980 epoch 1 - iter 2605/5212 - loss 0.49511139 - time (sec): 127.02 - samples/sec: 1439.63 - lr: 0.000025 - momentum: 0.000000
81
+ 2023-10-15 18:33:27,406 epoch 1 - iter 3126/5212 - loss 0.44378646 - time (sec): 153.44 - samples/sec: 1449.27 - lr: 0.000030 - momentum: 0.000000
82
+ 2023-10-15 18:33:52,795 epoch 1 - iter 3647/5212 - loss 0.41309918 - time (sec): 178.83 - samples/sec: 1447.91 - lr: 0.000035 - momentum: 0.000000
83
+ 2023-10-15 18:34:17,986 epoch 1 - iter 4168/5212 - loss 0.38982212 - time (sec): 204.02 - samples/sec: 1449.28 - lr: 0.000040 - momentum: 0.000000
84
+ 2023-10-15 18:34:43,243 epoch 1 - iter 4689/5212 - loss 0.37676637 - time (sec): 229.28 - samples/sec: 1435.39 - lr: 0.000045 - momentum: 0.000000
85
+ 2023-10-15 18:35:09,023 epoch 1 - iter 5210/5212 - loss 0.35899282 - time (sec): 255.06 - samples/sec: 1440.23 - lr: 0.000050 - momentum: 0.000000
86
+ 2023-10-15 18:35:09,114 ----------------------------------------------------------------------------------------------------
87
+ 2023-10-15 18:35:09,114 EPOCH 1 done: loss 0.3590 - lr: 0.000050
88
+ 2023-10-15 18:35:14,909 DEV : loss 0.1935647428035736 - f1-score (micro avg) 0.2561
89
+ 2023-10-15 18:35:14,934 saving best model
90
+ 2023-10-15 18:35:15,282 ----------------------------------------------------------------------------------------------------
91
+ 2023-10-15 18:35:40,738 epoch 2 - iter 521/5212 - loss 0.24636769 - time (sec): 25.45 - samples/sec: 1320.35 - lr: 0.000049 - momentum: 0.000000
92
+ 2023-10-15 18:36:05,977 epoch 2 - iter 1042/5212 - loss 0.23231567 - time (sec): 50.69 - samples/sec: 1396.46 - lr: 0.000049 - momentum: 0.000000
93
+ 2023-10-15 18:36:31,184 epoch 2 - iter 1563/5212 - loss 0.22395732 - time (sec): 75.90 - samples/sec: 1398.81 - lr: 0.000048 - momentum: 0.000000
94
+ 2023-10-15 18:36:56,597 epoch 2 - iter 2084/5212 - loss 0.21548760 - time (sec): 101.31 - samples/sec: 1415.13 - lr: 0.000048 - momentum: 0.000000
95
+ 2023-10-15 18:37:21,551 epoch 2 - iter 2605/5212 - loss 0.21176218 - time (sec): 126.27 - samples/sec: 1419.20 - lr: 0.000047 - momentum: 0.000000
96
+ 2023-10-15 18:37:47,270 epoch 2 - iter 3126/5212 - loss 0.20948142 - time (sec): 151.99 - samples/sec: 1438.50 - lr: 0.000047 - momentum: 0.000000
97
+ 2023-10-15 18:38:12,325 epoch 2 - iter 3647/5212 - loss 0.20703935 - time (sec): 177.04 - samples/sec: 1437.21 - lr: 0.000046 - momentum: 0.000000
98
+ 2023-10-15 18:38:38,051 epoch 2 - iter 4168/5212 - loss 0.20284934 - time (sec): 202.77 - samples/sec: 1435.09 - lr: 0.000046 - momentum: 0.000000
99
+ 2023-10-15 18:39:03,756 epoch 2 - iter 4689/5212 - loss 0.19868831 - time (sec): 228.47 - samples/sec: 1437.02 - lr: 0.000045 - momentum: 0.000000
100
+ 2023-10-15 18:39:29,272 epoch 2 - iter 5210/5212 - loss 0.19660262 - time (sec): 253.99 - samples/sec: 1444.41 - lr: 0.000044 - momentum: 0.000000
101
+ 2023-10-15 18:39:29,439 ----------------------------------------------------------------------------------------------------
102
+ 2023-10-15 18:39:29,439 EPOCH 2 done: loss 0.1966 - lr: 0.000044
103
+ 2023-10-15 18:39:38,418 DEV : loss 0.26646727323532104 - f1-score (micro avg) 0.3126
104
+ 2023-10-15 18:39:38,445 saving best model
105
+ 2023-10-15 18:39:38,908 ----------------------------------------------------------------------------------------------------
106
+ 2023-10-15 18:40:03,583 epoch 3 - iter 521/5212 - loss 0.15758472 - time (sec): 24.67 - samples/sec: 1359.66 - lr: 0.000044 - momentum: 0.000000
107
+ 2023-10-15 18:40:28,809 epoch 3 - iter 1042/5212 - loss 0.15326080 - time (sec): 49.90 - samples/sec: 1378.81 - lr: 0.000043 - momentum: 0.000000
108
+ 2023-10-15 18:40:55,000 epoch 3 - iter 1563/5212 - loss 0.14597497 - time (sec): 76.09 - samples/sec: 1452.84 - lr: 0.000043 - momentum: 0.000000
109
+ 2023-10-15 18:41:20,559 epoch 3 - iter 2084/5212 - loss 0.14065920 - time (sec): 101.65 - samples/sec: 1454.24 - lr: 0.000042 - momentum: 0.000000
110
+ 2023-10-15 18:41:45,973 epoch 3 - iter 2605/5212 - loss 0.14592720 - time (sec): 127.06 - samples/sec: 1454.39 - lr: 0.000042 - momentum: 0.000000
111
+ 2023-10-15 18:42:10,959 epoch 3 - iter 3126/5212 - loss 0.14605254 - time (sec): 152.05 - samples/sec: 1458.06 - lr: 0.000041 - momentum: 0.000000
112
+ 2023-10-15 18:42:36,131 epoch 3 - iter 3647/5212 - loss 0.14458467 - time (sec): 177.22 - samples/sec: 1455.44 - lr: 0.000041 - momentum: 0.000000
113
+ 2023-10-15 18:43:01,039 epoch 3 - iter 4168/5212 - loss 0.14831744 - time (sec): 202.13 - samples/sec: 1448.82 - lr: 0.000040 - momentum: 0.000000
114
+ 2023-10-15 18:43:26,000 epoch 3 - iter 4689/5212 - loss 0.14722580 - time (sec): 227.09 - samples/sec: 1449.99 - lr: 0.000039 - momentum: 0.000000
115
+ 2023-10-15 18:43:51,506 epoch 3 - iter 5210/5212 - loss 0.14520606 - time (sec): 252.60 - samples/sec: 1454.29 - lr: 0.000039 - momentum: 0.000000
116
+ 2023-10-15 18:43:51,594 ----------------------------------------------------------------------------------------------------
117
+ 2023-10-15 18:43:51,594 EPOCH 3 done: loss 0.1452 - lr: 0.000039
118
+ 2023-10-15 18:43:59,942 DEV : loss 0.21102237701416016 - f1-score (micro avg) 0.3177
119
+ 2023-10-15 18:43:59,971 saving best model
120
+ 2023-10-15 18:44:00,404 ----------------------------------------------------------------------------------------------------
121
+ 2023-10-15 18:44:26,791 epoch 4 - iter 521/5212 - loss 0.09694449 - time (sec): 26.39 - samples/sec: 1405.34 - lr: 0.000038 - momentum: 0.000000
122
+ 2023-10-15 18:44:51,970 epoch 4 - iter 1042/5212 - loss 0.10160415 - time (sec): 51.56 - samples/sec: 1458.36 - lr: 0.000038 - momentum: 0.000000
123
+ 2023-10-15 18:45:17,568 epoch 4 - iter 1563/5212 - loss 0.10470997 - time (sec): 77.16 - samples/sec: 1440.66 - lr: 0.000037 - momentum: 0.000000
124
+ 2023-10-15 18:45:42,943 epoch 4 - iter 2084/5212 - loss 0.10504412 - time (sec): 102.54 - samples/sec: 1437.98 - lr: 0.000037 - momentum: 0.000000
125
+ 2023-10-15 18:46:08,464 epoch 4 - iter 2605/5212 - loss 0.10205242 - time (sec): 128.06 - samples/sec: 1436.72 - lr: 0.000036 - momentum: 0.000000
126
+ 2023-10-15 18:46:34,112 epoch 4 - iter 3126/5212 - loss 0.10491793 - time (sec): 153.71 - samples/sec: 1442.53 - lr: 0.000036 - momentum: 0.000000
127
+ 2023-10-15 18:46:59,262 epoch 4 - iter 3647/5212 - loss 0.10491333 - time (sec): 178.86 - samples/sec: 1451.17 - lr: 0.000035 - momentum: 0.000000
128
+ 2023-10-15 18:47:24,773 epoch 4 - iter 4168/5212 - loss 0.11104350 - time (sec): 204.37 - samples/sec: 1443.22 - lr: 0.000034 - momentum: 0.000000
129
+ 2023-10-15 18:47:49,932 epoch 4 - iter 4689/5212 - loss 0.11023634 - time (sec): 229.53 - samples/sec: 1438.50 - lr: 0.000034 - momentum: 0.000000
130
+ 2023-10-15 18:48:15,080 epoch 4 - iter 5210/5212 - loss 0.11005524 - time (sec): 254.67 - samples/sec: 1442.59 - lr: 0.000033 - momentum: 0.000000
131
+ 2023-10-15 18:48:15,165 ----------------------------------------------------------------------------------------------------
132
+ 2023-10-15 18:48:15,165 EPOCH 4 done: loss 0.1100 - lr: 0.000033
133
+ 2023-10-15 18:48:23,650 DEV : loss 0.3510819673538208 - f1-score (micro avg) 0.3189
134
+ 2023-10-15 18:48:23,695 saving best model
135
+ 2023-10-15 18:48:24,183 ----------------------------------------------------------------------------------------------------
136
+ 2023-10-15 18:48:51,942 epoch 5 - iter 521/5212 - loss 0.08171099 - time (sec): 27.76 - samples/sec: 1437.83 - lr: 0.000033 - momentum: 0.000000
137
+ 2023-10-15 18:49:17,391 epoch 5 - iter 1042/5212 - loss 0.07891529 - time (sec): 53.21 - samples/sec: 1429.22 - lr: 0.000032 - momentum: 0.000000
138
+ 2023-10-15 18:49:43,253 epoch 5 - iter 1563/5212 - loss 0.08091614 - time (sec): 79.07 - samples/sec: 1415.21 - lr: 0.000032 - momentum: 0.000000
139
+ 2023-10-15 18:50:08,706 epoch 5 - iter 2084/5212 - loss 0.07786060 - time (sec): 104.52 - samples/sec: 1436.25 - lr: 0.000031 - momentum: 0.000000
140
+ 2023-10-15 18:50:33,799 epoch 5 - iter 2605/5212 - loss 0.07747685 - time (sec): 129.61 - samples/sec: 1443.70 - lr: 0.000031 - momentum: 0.000000
141
+ 2023-10-15 18:50:58,814 epoch 5 - iter 3126/5212 - loss 0.07663910 - time (sec): 154.63 - samples/sec: 1438.00 - lr: 0.000030 - momentum: 0.000000
142
+ 2023-10-15 18:51:23,868 epoch 5 - iter 3647/5212 - loss 0.07608571 - time (sec): 179.68 - samples/sec: 1428.61 - lr: 0.000029 - momentum: 0.000000
143
+ 2023-10-15 18:51:49,382 epoch 5 - iter 4168/5212 - loss 0.07640161 - time (sec): 205.20 - samples/sec: 1432.32 - lr: 0.000029 - momentum: 0.000000
144
+ 2023-10-15 18:52:14,940 epoch 5 - iter 4689/5212 - loss 0.07755494 - time (sec): 230.76 - samples/sec: 1433.82 - lr: 0.000028 - momentum: 0.000000
145
+ 2023-10-15 18:52:39,947 epoch 5 - iter 5210/5212 - loss 0.07842196 - time (sec): 255.76 - samples/sec: 1436.38 - lr: 0.000028 - momentum: 0.000000
146
+ 2023-10-15 18:52:40,037 ----------------------------------------------------------------------------------------------------
147
+ 2023-10-15 18:52:40,037 EPOCH 5 done: loss 0.0784 - lr: 0.000028
148
+ 2023-10-15 18:52:48,525 DEV : loss 0.2648324966430664 - f1-score (micro avg) 0.3525
149
+ 2023-10-15 18:52:48,571 saving best model
150
+ 2023-10-15 18:52:49,098 ----------------------------------------------------------------------------------------------------
151
+ 2023-10-15 18:53:15,489 epoch 6 - iter 521/5212 - loss 0.05230408 - time (sec): 26.39 - samples/sec: 1360.55 - lr: 0.000027 - momentum: 0.000000
152
+ 2023-10-15 18:53:40,907 epoch 6 - iter 1042/5212 - loss 0.05845790 - time (sec): 51.81 - samples/sec: 1434.46 - lr: 0.000027 - momentum: 0.000000
153
+ 2023-10-15 18:54:07,100 epoch 6 - iter 1563/5212 - loss 0.06310501 - time (sec): 78.00 - samples/sec: 1439.16 - lr: 0.000026 - momentum: 0.000000
154
+ 2023-10-15 18:54:32,649 epoch 6 - iter 2084/5212 - loss 0.06369300 - time (sec): 103.55 - samples/sec: 1433.91 - lr: 0.000026 - momentum: 0.000000
155
+ 2023-10-15 18:54:57,100 epoch 6 - iter 2605/5212 - loss 0.06667131 - time (sec): 128.00 - samples/sec: 1423.71 - lr: 0.000025 - momentum: 0.000000
156
+ 2023-10-15 18:55:22,500 epoch 6 - iter 3126/5212 - loss 0.06631037 - time (sec): 153.40 - samples/sec: 1441.54 - lr: 0.000024 - momentum: 0.000000
157
+ 2023-10-15 18:55:48,175 epoch 6 - iter 3647/5212 - loss 0.06540271 - time (sec): 179.07 - samples/sec: 1434.50 - lr: 0.000024 - momentum: 0.000000
158
+ 2023-10-15 18:56:13,599 epoch 6 - iter 4168/5212 - loss 0.06511901 - time (sec): 204.50 - samples/sec: 1439.13 - lr: 0.000023 - momentum: 0.000000
159
+ 2023-10-15 18:56:38,437 epoch 6 - iter 4689/5212 - loss 0.06326195 - time (sec): 229.34 - samples/sec: 1443.32 - lr: 0.000023 - momentum: 0.000000
160
+ 2023-10-15 18:57:03,252 epoch 6 - iter 5210/5212 - loss 0.06322970 - time (sec): 254.15 - samples/sec: 1444.22 - lr: 0.000022 - momentum: 0.000000
161
+ 2023-10-15 18:57:03,445 ----------------------------------------------------------------------------------------------------
162
+ 2023-10-15 18:57:03,446 EPOCH 6 done: loss 0.0632 - lr: 0.000022
163
+ 2023-10-15 18:57:11,710 DEV : loss 0.3284428119659424 - f1-score (micro avg) 0.3479
164
+ 2023-10-15 18:57:11,739 ----------------------------------------------------------------------------------------------------
165
+ 2023-10-15 18:57:36,385 epoch 7 - iter 521/5212 - loss 0.04602036 - time (sec): 24.65 - samples/sec: 1375.30 - lr: 0.000022 - momentum: 0.000000
166
+ 2023-10-15 18:58:01,491 epoch 7 - iter 1042/5212 - loss 0.04163037 - time (sec): 49.75 - samples/sec: 1426.66 - lr: 0.000021 - momentum: 0.000000
167
+ 2023-10-15 18:58:26,685 epoch 7 - iter 1563/5212 - loss 0.04302604 - time (sec): 74.94 - samples/sec: 1400.93 - lr: 0.000021 - momentum: 0.000000
168
+ 2023-10-15 18:58:52,373 epoch 7 - iter 2084/5212 - loss 0.04090613 - time (sec): 100.63 - samples/sec: 1421.62 - lr: 0.000020 - momentum: 0.000000
169
+ 2023-10-15 18:59:18,332 epoch 7 - iter 2605/5212 - loss 0.04297160 - time (sec): 126.59 - samples/sec: 1428.63 - lr: 0.000019 - momentum: 0.000000
170
+ 2023-10-15 18:59:43,473 epoch 7 - iter 3126/5212 - loss 0.04222747 - time (sec): 151.73 - samples/sec: 1430.24 - lr: 0.000019 - momentum: 0.000000
171
+ 2023-10-15 19:00:08,729 epoch 7 - iter 3647/5212 - loss 0.04304694 - time (sec): 176.99 - samples/sec: 1441.97 - lr: 0.000018 - momentum: 0.000000
172
+ 2023-10-15 19:00:33,766 epoch 7 - iter 4168/5212 - loss 0.04368892 - time (sec): 202.03 - samples/sec: 1449.26 - lr: 0.000018 - momentum: 0.000000
173
+ 2023-10-15 19:00:59,308 epoch 7 - iter 4689/5212 - loss 0.04297916 - time (sec): 227.57 - samples/sec: 1447.31 - lr: 0.000017 - momentum: 0.000000
174
+ 2023-10-15 19:01:24,566 epoch 7 - iter 5210/5212 - loss 0.04365791 - time (sec): 252.83 - samples/sec: 1453.14 - lr: 0.000017 - momentum: 0.000000
175
+ 2023-10-15 19:01:24,659 ----------------------------------------------------------------------------------------------------
176
+ 2023-10-15 19:01:24,659 EPOCH 7 done: loss 0.0437 - lr: 0.000017
177
+ 2023-10-15 19:01:33,790 DEV : loss 0.3804240822792053 - f1-score (micro avg) 0.351
178
+ 2023-10-15 19:01:33,821 ----------------------------------------------------------------------------------------------------
179
+ 2023-10-15 19:01:59,298 epoch 8 - iter 521/5212 - loss 0.03278593 - time (sec): 25.48 - samples/sec: 1544.77 - lr: 0.000016 - momentum: 0.000000
180
+ 2023-10-15 19:02:25,536 epoch 8 - iter 1042/5212 - loss 0.03015265 - time (sec): 51.71 - samples/sec: 1503.92 - lr: 0.000016 - momentum: 0.000000
181
+ 2023-10-15 19:02:52,003 epoch 8 - iter 1563/5212 - loss 0.03141214 - time (sec): 78.18 - samples/sec: 1470.41 - lr: 0.000015 - momentum: 0.000000
182
+ 2023-10-15 19:03:17,168 epoch 8 - iter 2084/5212 - loss 0.03397435 - time (sec): 103.35 - samples/sec: 1450.36 - lr: 0.000014 - momentum: 0.000000
183
+ 2023-10-15 19:03:42,358 epoch 8 - iter 2605/5212 - loss 0.03614872 - time (sec): 128.54 - samples/sec: 1440.92 - lr: 0.000014 - momentum: 0.000000
184
+ 2023-10-15 19:04:08,045 epoch 8 - iter 3126/5212 - loss 0.03527472 - time (sec): 154.22 - samples/sec: 1436.27 - lr: 0.000013 - momentum: 0.000000
185
+ 2023-10-15 19:04:34,122 epoch 8 - iter 3647/5212 - loss 0.03459124 - time (sec): 180.30 - samples/sec: 1439.85 - lr: 0.000013 - momentum: 0.000000
186
+ 2023-10-15 19:04:58,912 epoch 8 - iter 4168/5212 - loss 0.03421361 - time (sec): 205.09 - samples/sec: 1434.35 - lr: 0.000012 - momentum: 0.000000
187
+ 2023-10-15 19:05:23,753 epoch 8 - iter 4689/5212 - loss 0.03359174 - time (sec): 229.93 - samples/sec: 1436.19 - lr: 0.000012 - momentum: 0.000000
188
+ 2023-10-15 19:05:48,530 epoch 8 - iter 5210/5212 - loss 0.03303597 - time (sec): 254.71 - samples/sec: 1440.90 - lr: 0.000011 - momentum: 0.000000
189
+ 2023-10-15 19:05:48,643 ----------------------------------------------------------------------------------------------------
190
+ 2023-10-15 19:05:48,643 EPOCH 8 done: loss 0.0330 - lr: 0.000011
191
+ 2023-10-15 19:05:58,972 DEV : loss 0.3673393130302429 - f1-score (micro avg) 0.3845
192
+ 2023-10-15 19:05:59,018 saving best model
193
+ 2023-10-15 19:05:59,528 ----------------------------------------------------------------------------------------------------
194
+ 2023-10-15 19:06:28,356 epoch 9 - iter 521/5212 - loss 0.02522660 - time (sec): 28.82 - samples/sec: 1234.63 - lr: 0.000011 - momentum: 0.000000
195
+ 2023-10-15 19:06:55,424 epoch 9 - iter 1042/5212 - loss 0.02383534 - time (sec): 55.89 - samples/sec: 1282.26 - lr: 0.000010 - momentum: 0.000000
196
+ 2023-10-15 19:07:20,832 epoch 9 - iter 1563/5212 - loss 0.02195053 - time (sec): 81.30 - samples/sec: 1331.42 - lr: 0.000009 - momentum: 0.000000
197
+ 2023-10-15 19:07:45,994 epoch 9 - iter 2084/5212 - loss 0.02261221 - time (sec): 106.46 - samples/sec: 1360.19 - lr: 0.000009 - momentum: 0.000000
198
+ 2023-10-15 19:08:11,005 epoch 9 - iter 2605/5212 - loss 0.02502914 - time (sec): 131.47 - samples/sec: 1370.83 - lr: 0.000008 - momentum: 0.000000
199
+ 2023-10-15 19:08:37,062 epoch 9 - iter 3126/5212 - loss 0.02521692 - time (sec): 157.53 - samples/sec: 1390.80 - lr: 0.000008 - momentum: 0.000000
200
+ 2023-10-15 19:09:02,395 epoch 9 - iter 3647/5212 - loss 0.02518572 - time (sec): 182.86 - samples/sec: 1397.05 - lr: 0.000007 - momentum: 0.000000
201
+ 2023-10-15 19:09:27,771 epoch 9 - iter 4168/5212 - loss 0.02436796 - time (sec): 208.24 - samples/sec: 1400.34 - lr: 0.000007 - momentum: 0.000000
202
+ 2023-10-15 19:09:53,231 epoch 9 - iter 4689/5212 - loss 0.02383592 - time (sec): 233.70 - samples/sec: 1411.35 - lr: 0.000006 - momentum: 0.000000
203
+ 2023-10-15 19:10:18,608 epoch 9 - iter 5210/5212 - loss 0.02324373 - time (sec): 259.08 - samples/sec: 1417.78 - lr: 0.000006 - momentum: 0.000000
204
+ 2023-10-15 19:10:18,697 ----------------------------------------------------------------------------------------------------
205
+ 2023-10-15 19:10:18,697 EPOCH 9 done: loss 0.0232 - lr: 0.000006
206
+ 2023-10-15 19:10:28,083 DEV : loss 0.4655354917049408 - f1-score (micro avg) 0.3527
207
+ 2023-10-15 19:10:28,119 ----------------------------------------------------------------------------------------------------
208
+ 2023-10-15 19:10:55,276 epoch 10 - iter 521/5212 - loss 0.02072785 - time (sec): 27.16 - samples/sec: 1322.25 - lr: 0.000005 - momentum: 0.000000
209
+ 2023-10-15 19:11:20,967 epoch 10 - iter 1042/5212 - loss 0.01880289 - time (sec): 52.85 - samples/sec: 1386.39 - lr: 0.000004 - momentum: 0.000000
210
+ 2023-10-15 19:11:46,669 epoch 10 - iter 1563/5212 - loss 0.01711965 - time (sec): 78.55 - samples/sec: 1407.78 - lr: 0.000004 - momentum: 0.000000
211
+ 2023-10-15 19:12:11,914 epoch 10 - iter 2084/5212 - loss 0.01667270 - time (sec): 103.79 - samples/sec: 1416.99 - lr: 0.000003 - momentum: 0.000000
212
+ 2023-10-15 19:12:37,387 epoch 10 - iter 2605/5212 - loss 0.01655129 - time (sec): 129.27 - samples/sec: 1436.27 - lr: 0.000003 - momentum: 0.000000
213
+ 2023-10-15 19:13:03,491 epoch 10 - iter 3126/5212 - loss 0.01639347 - time (sec): 155.37 - samples/sec: 1430.40 - lr: 0.000002 - momentum: 0.000000
214
+ 2023-10-15 19:13:29,109 epoch 10 - iter 3647/5212 - loss 0.01559952 - time (sec): 180.99 - samples/sec: 1430.74 - lr: 0.000002 - momentum: 0.000000
215
+ 2023-10-15 19:13:54,015 epoch 10 - iter 4168/5212 - loss 0.01527630 - time (sec): 205.89 - samples/sec: 1427.16 - lr: 0.000001 - momentum: 0.000000
216
+ 2023-10-15 19:14:19,322 epoch 10 - iter 4689/5212 - loss 0.01494606 - time (sec): 231.20 - samples/sec: 1425.82 - lr: 0.000001 - momentum: 0.000000
217
+ 2023-10-15 19:14:44,781 epoch 10 - iter 5210/5212 - loss 0.01478709 - time (sec): 256.66 - samples/sec: 1430.28 - lr: 0.000000 - momentum: 0.000000
218
+ 2023-10-15 19:14:44,923 ----------------------------------------------------------------------------------------------------
219
+ 2023-10-15 19:14:44,923 EPOCH 10 done: loss 0.0148 - lr: 0.000000
220
+ 2023-10-15 19:14:53,969 DEV : loss 0.4225836992263794 - f1-score (micro avg) 0.3731
221
+ 2023-10-15 19:14:54,376 ----------------------------------------------------------------------------------------------------
222
+ 2023-10-15 19:14:54,377 Loading model from best epoch ...
223
+ 2023-10-15 19:14:55,998 SequenceTagger predicts: Dictionary with 17 tags: O, S-LOC, B-LOC, E-LOC, I-LOC, S-PER, B-PER, E-PER, I-PER, S-ORG, B-ORG, E-ORG, I-ORG, S-HumanProd, B-HumanProd, E-HumanProd, I-HumanProd
224
+ 2023-10-15 19:15:12,852
225
+ Results:
226
+ - F-score (micro) 0.4487
227
+ - F-score (macro) 0.2891
228
+ - Accuracy 0.2932
229
+
230
+ By class:
231
+ precision recall f1-score support
232
+
233
+ LOC 0.5092 0.5708 0.5383 1214
234
+ PER 0.3792 0.4059 0.3921 808
235
+ ORG 0.2928 0.1841 0.2261 353
236
+ HumanProd 0.0000 0.0000 0.0000 15
237
+
238
+ micro avg 0.4431 0.4544 0.4487 2390
239
+ macro avg 0.2953 0.2902 0.2891 2390
240
+ weighted avg 0.4301 0.4544 0.4394 2390
241
+
242
+ 2023-10-15 19:15:12,852 ----------------------------------------------------------------------------------------------------