File size: 24,225 Bytes
a41d7ec |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 |
2023-10-17 10:23:49,346 ----------------------------------------------------------------------------------------------------
2023-10-17 10:23:49,348 Model: "SequenceTagger(
(embeddings): TransformerWordEmbeddings(
(model): ElectraModel(
(embeddings): ElectraEmbeddings(
(word_embeddings): Embedding(32001, 768)
(position_embeddings): Embedding(512, 768)
(token_type_embeddings): Embedding(2, 768)
(LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
(dropout): Dropout(p=0.1, inplace=False)
)
(encoder): ElectraEncoder(
(layer): ModuleList(
(0-11): 12 x ElectraLayer(
(attention): ElectraAttention(
(self): ElectraSelfAttention(
(query): Linear(in_features=768, out_features=768, bias=True)
(key): Linear(in_features=768, out_features=768, bias=True)
(value): Linear(in_features=768, out_features=768, bias=True)
(dropout): Dropout(p=0.1, inplace=False)
)
(output): ElectraSelfOutput(
(dense): Linear(in_features=768, out_features=768, bias=True)
(LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
(dropout): Dropout(p=0.1, inplace=False)
)
)
(intermediate): ElectraIntermediate(
(dense): Linear(in_features=768, out_features=3072, bias=True)
(intermediate_act_fn): GELUActivation()
)
(output): ElectraOutput(
(dense): Linear(in_features=3072, out_features=768, bias=True)
(LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
(dropout): Dropout(p=0.1, inplace=False)
)
)
)
)
)
)
(locked_dropout): LockedDropout(p=0.5)
(linear): Linear(in_features=768, out_features=13, bias=True)
(loss_function): CrossEntropyLoss()
)"
2023-10-17 10:23:49,348 ----------------------------------------------------------------------------------------------------
2023-10-17 10:23:49,348 MultiCorpus: 14465 train + 1392 dev + 2432 test sentences
- NER_HIPE_2022 Corpus: 14465 train + 1392 dev + 2432 test sentences - /root/.flair/datasets/ner_hipe_2022/v2.1/letemps/fr/with_doc_seperator
2023-10-17 10:23:49,348 ----------------------------------------------------------------------------------------------------
2023-10-17 10:23:49,348 Train: 14465 sentences
2023-10-17 10:23:49,348 (train_with_dev=False, train_with_test=False)
2023-10-17 10:23:49,348 ----------------------------------------------------------------------------------------------------
2023-10-17 10:23:49,348 Training Params:
2023-10-17 10:23:49,348 - learning_rate: "3e-05"
2023-10-17 10:23:49,348 - mini_batch_size: "8"
2023-10-17 10:23:49,348 - max_epochs: "10"
2023-10-17 10:23:49,348 - shuffle: "True"
2023-10-17 10:23:49,348 ----------------------------------------------------------------------------------------------------
2023-10-17 10:23:49,348 Plugins:
2023-10-17 10:23:49,348 - TensorboardLogger
2023-10-17 10:23:49,348 - LinearScheduler | warmup_fraction: '0.1'
2023-10-17 10:23:49,348 ----------------------------------------------------------------------------------------------------
2023-10-17 10:23:49,348 Final evaluation on model from best epoch (best-model.pt)
2023-10-17 10:23:49,348 - metric: "('micro avg', 'f1-score')"
2023-10-17 10:23:49,348 ----------------------------------------------------------------------------------------------------
2023-10-17 10:23:49,349 Computation:
2023-10-17 10:23:49,349 - compute on device: cuda:0
2023-10-17 10:23:49,349 - embedding storage: none
2023-10-17 10:23:49,349 ----------------------------------------------------------------------------------------------------
2023-10-17 10:23:49,349 Model training base path: "hmbench-letemps/fr-hmteams/teams-base-historic-multilingual-discriminator-bs8-wsFalse-e10-lr3e-05-poolingfirst-layers-1-crfFalse-1"
2023-10-17 10:23:49,349 ----------------------------------------------------------------------------------------------------
2023-10-17 10:23:49,349 ----------------------------------------------------------------------------------------------------
2023-10-17 10:23:49,349 Logging anything other than scalars to TensorBoard is currently not supported.
2023-10-17 10:24:02,854 epoch 1 - iter 180/1809 - loss 1.97077795 - time (sec): 13.50 - samples/sec: 2869.63 - lr: 0.000003 - momentum: 0.000000
2023-10-17 10:24:15,874 epoch 1 - iter 360/1809 - loss 1.13419673 - time (sec): 26.52 - samples/sec: 2872.06 - lr: 0.000006 - momentum: 0.000000
2023-10-17 10:24:29,572 epoch 1 - iter 540/1809 - loss 0.80989982 - time (sec): 40.22 - samples/sec: 2841.00 - lr: 0.000009 - momentum: 0.000000
2023-10-17 10:24:43,353 epoch 1 - iter 720/1809 - loss 0.64354116 - time (sec): 54.00 - samples/sec: 2825.49 - lr: 0.000012 - momentum: 0.000000
2023-10-17 10:24:57,139 epoch 1 - iter 900/1809 - loss 0.54278891 - time (sec): 67.79 - samples/sec: 2798.57 - lr: 0.000015 - momentum: 0.000000
2023-10-17 10:25:11,974 epoch 1 - iter 1080/1809 - loss 0.47491789 - time (sec): 82.62 - samples/sec: 2741.90 - lr: 0.000018 - momentum: 0.000000
2023-10-17 10:25:26,384 epoch 1 - iter 1260/1809 - loss 0.42566960 - time (sec): 97.03 - samples/sec: 2718.52 - lr: 0.000021 - momentum: 0.000000
2023-10-17 10:25:41,389 epoch 1 - iter 1440/1809 - loss 0.38559522 - time (sec): 112.04 - samples/sec: 2704.90 - lr: 0.000024 - momentum: 0.000000
2023-10-17 10:25:54,544 epoch 1 - iter 1620/1809 - loss 0.35386098 - time (sec): 125.19 - samples/sec: 2725.72 - lr: 0.000027 - momentum: 0.000000
2023-10-17 10:26:07,351 epoch 1 - iter 1800/1809 - loss 0.32934241 - time (sec): 138.00 - samples/sec: 2738.94 - lr: 0.000030 - momentum: 0.000000
2023-10-17 10:26:08,063 ----------------------------------------------------------------------------------------------------
2023-10-17 10:26:08,063 EPOCH 1 done: loss 0.3284 - lr: 0.000030
2023-10-17 10:26:14,169 DEV : loss 0.09768907725811005 - f1-score (micro avg) 0.6108
2023-10-17 10:26:14,210 saving best model
2023-10-17 10:26:14,712 ----------------------------------------------------------------------------------------------------
2023-10-17 10:26:27,936 epoch 2 - iter 180/1809 - loss 0.10126158 - time (sec): 13.22 - samples/sec: 2761.32 - lr: 0.000030 - momentum: 0.000000
2023-10-17 10:26:41,008 epoch 2 - iter 360/1809 - loss 0.09521427 - time (sec): 26.29 - samples/sec: 2883.92 - lr: 0.000029 - momentum: 0.000000
2023-10-17 10:26:54,116 epoch 2 - iter 540/1809 - loss 0.09155082 - time (sec): 39.40 - samples/sec: 2914.13 - lr: 0.000029 - momentum: 0.000000
2023-10-17 10:27:07,181 epoch 2 - iter 720/1809 - loss 0.09337842 - time (sec): 52.47 - samples/sec: 2885.06 - lr: 0.000029 - momentum: 0.000000
2023-10-17 10:27:21,201 epoch 2 - iter 900/1809 - loss 0.09093030 - time (sec): 66.49 - samples/sec: 2821.99 - lr: 0.000028 - momentum: 0.000000
2023-10-17 10:27:34,922 epoch 2 - iter 1080/1809 - loss 0.08938417 - time (sec): 80.21 - samples/sec: 2806.14 - lr: 0.000028 - momentum: 0.000000
2023-10-17 10:27:48,461 epoch 2 - iter 1260/1809 - loss 0.08853541 - time (sec): 93.75 - samples/sec: 2802.82 - lr: 0.000028 - momentum: 0.000000
2023-10-17 10:28:02,067 epoch 2 - iter 1440/1809 - loss 0.08910538 - time (sec): 107.35 - samples/sec: 2806.40 - lr: 0.000027 - momentum: 0.000000
2023-10-17 10:28:15,988 epoch 2 - iter 1620/1809 - loss 0.08699630 - time (sec): 121.27 - samples/sec: 2817.16 - lr: 0.000027 - momentum: 0.000000
2023-10-17 10:28:29,006 epoch 2 - iter 1800/1809 - loss 0.08639068 - time (sec): 134.29 - samples/sec: 2816.30 - lr: 0.000027 - momentum: 0.000000
2023-10-17 10:28:29,643 ----------------------------------------------------------------------------------------------------
2023-10-17 10:28:29,644 EPOCH 2 done: loss 0.0866 - lr: 0.000027
2023-10-17 10:28:36,718 DEV : loss 0.10609198361635208 - f1-score (micro avg) 0.6388
2023-10-17 10:28:36,759 saving best model
2023-10-17 10:28:37,363 ----------------------------------------------------------------------------------------------------
2023-10-17 10:28:52,654 epoch 3 - iter 180/1809 - loss 0.06507567 - time (sec): 15.29 - samples/sec: 2455.09 - lr: 0.000026 - momentum: 0.000000
2023-10-17 10:29:07,332 epoch 3 - iter 360/1809 - loss 0.06327625 - time (sec): 29.97 - samples/sec: 2571.73 - lr: 0.000026 - momentum: 0.000000
2023-10-17 10:29:22,132 epoch 3 - iter 540/1809 - loss 0.06394436 - time (sec): 44.77 - samples/sec: 2564.70 - lr: 0.000026 - momentum: 0.000000
2023-10-17 10:29:36,446 epoch 3 - iter 720/1809 - loss 0.06297439 - time (sec): 59.08 - samples/sec: 2604.56 - lr: 0.000025 - momentum: 0.000000
2023-10-17 10:29:49,726 epoch 3 - iter 900/1809 - loss 0.06209821 - time (sec): 72.36 - samples/sec: 2647.74 - lr: 0.000025 - momentum: 0.000000
2023-10-17 10:30:02,972 epoch 3 - iter 1080/1809 - loss 0.06128483 - time (sec): 85.61 - samples/sec: 2688.49 - lr: 0.000025 - momentum: 0.000000
2023-10-17 10:30:16,883 epoch 3 - iter 1260/1809 - loss 0.06138333 - time (sec): 99.52 - samples/sec: 2691.44 - lr: 0.000024 - momentum: 0.000000
2023-10-17 10:30:31,017 epoch 3 - iter 1440/1809 - loss 0.06203921 - time (sec): 113.65 - samples/sec: 2679.12 - lr: 0.000024 - momentum: 0.000000
2023-10-17 10:30:43,946 epoch 3 - iter 1620/1809 - loss 0.06242292 - time (sec): 126.58 - samples/sec: 2700.92 - lr: 0.000024 - momentum: 0.000000
2023-10-17 10:30:55,919 epoch 3 - iter 1800/1809 - loss 0.06256995 - time (sec): 138.55 - samples/sec: 2729.61 - lr: 0.000023 - momentum: 0.000000
2023-10-17 10:30:56,446 ----------------------------------------------------------------------------------------------------
2023-10-17 10:30:56,446 EPOCH 3 done: loss 0.0626 - lr: 0.000023
2023-10-17 10:31:02,865 DEV : loss 0.12596164643764496 - f1-score (micro avg) 0.6405
2023-10-17 10:31:02,910 saving best model
2023-10-17 10:31:03,530 ----------------------------------------------------------------------------------------------------
2023-10-17 10:31:17,877 epoch 4 - iter 180/1809 - loss 0.04168279 - time (sec): 14.34 - samples/sec: 2707.50 - lr: 0.000023 - momentum: 0.000000
2023-10-17 10:31:31,828 epoch 4 - iter 360/1809 - loss 0.03962130 - time (sec): 28.30 - samples/sec: 2707.68 - lr: 0.000023 - momentum: 0.000000
2023-10-17 10:31:46,220 epoch 4 - iter 540/1809 - loss 0.04238080 - time (sec): 42.69 - samples/sec: 2714.96 - lr: 0.000022 - momentum: 0.000000
2023-10-17 10:31:59,189 epoch 4 - iter 720/1809 - loss 0.04283943 - time (sec): 55.66 - samples/sec: 2743.08 - lr: 0.000022 - momentum: 0.000000
2023-10-17 10:32:12,306 epoch 4 - iter 900/1809 - loss 0.04535463 - time (sec): 68.77 - samples/sec: 2782.02 - lr: 0.000022 - momentum: 0.000000
2023-10-17 10:32:25,119 epoch 4 - iter 1080/1809 - loss 0.04573226 - time (sec): 81.59 - samples/sec: 2784.61 - lr: 0.000021 - momentum: 0.000000
2023-10-17 10:32:38,107 epoch 4 - iter 1260/1809 - loss 0.04529284 - time (sec): 94.57 - samples/sec: 2803.28 - lr: 0.000021 - momentum: 0.000000
2023-10-17 10:32:51,646 epoch 4 - iter 1440/1809 - loss 0.04597535 - time (sec): 108.11 - samples/sec: 2807.64 - lr: 0.000021 - momentum: 0.000000
2023-10-17 10:33:05,014 epoch 4 - iter 1620/1809 - loss 0.04539057 - time (sec): 121.48 - samples/sec: 2801.07 - lr: 0.000020 - momentum: 0.000000
2023-10-17 10:33:18,630 epoch 4 - iter 1800/1809 - loss 0.04608749 - time (sec): 135.10 - samples/sec: 2798.18 - lr: 0.000020 - momentum: 0.000000
2023-10-17 10:33:19,371 ----------------------------------------------------------------------------------------------------
2023-10-17 10:33:19,371 EPOCH 4 done: loss 0.0461 - lr: 0.000020
2023-10-17 10:33:26,556 DEV : loss 0.17203289270401 - f1-score (micro avg) 0.6513
2023-10-17 10:33:26,598 saving best model
2023-10-17 10:33:27,228 ----------------------------------------------------------------------------------------------------
2023-10-17 10:33:41,243 epoch 5 - iter 180/1809 - loss 0.02768648 - time (sec): 14.01 - samples/sec: 2708.81 - lr: 0.000020 - momentum: 0.000000
2023-10-17 10:33:54,303 epoch 5 - iter 360/1809 - loss 0.03102599 - time (sec): 27.07 - samples/sec: 2738.17 - lr: 0.000019 - momentum: 0.000000
2023-10-17 10:34:08,359 epoch 5 - iter 540/1809 - loss 0.03185567 - time (sec): 41.13 - samples/sec: 2734.73 - lr: 0.000019 - momentum: 0.000000
2023-10-17 10:34:21,848 epoch 5 - iter 720/1809 - loss 0.03061780 - time (sec): 54.62 - samples/sec: 2737.94 - lr: 0.000019 - momentum: 0.000000
2023-10-17 10:34:36,525 epoch 5 - iter 900/1809 - loss 0.03119659 - time (sec): 69.29 - samples/sec: 2723.36 - lr: 0.000018 - momentum: 0.000000
2023-10-17 10:34:50,562 epoch 5 - iter 1080/1809 - loss 0.03101663 - time (sec): 83.33 - samples/sec: 2720.38 - lr: 0.000018 - momentum: 0.000000
2023-10-17 10:35:04,954 epoch 5 - iter 1260/1809 - loss 0.03070119 - time (sec): 97.72 - samples/sec: 2718.56 - lr: 0.000018 - momentum: 0.000000
2023-10-17 10:35:19,209 epoch 5 - iter 1440/1809 - loss 0.03061179 - time (sec): 111.98 - samples/sec: 2706.42 - lr: 0.000017 - momentum: 0.000000
2023-10-17 10:35:33,200 epoch 5 - iter 1620/1809 - loss 0.03213924 - time (sec): 125.97 - samples/sec: 2700.86 - lr: 0.000017 - momentum: 0.000000
2023-10-17 10:35:46,610 epoch 5 - iter 1800/1809 - loss 0.03232409 - time (sec): 139.38 - samples/sec: 2712.26 - lr: 0.000017 - momentum: 0.000000
2023-10-17 10:35:47,324 ----------------------------------------------------------------------------------------------------
2023-10-17 10:35:47,324 EPOCH 5 done: loss 0.0324 - lr: 0.000017
2023-10-17 10:35:53,774 DEV : loss 0.2933090031147003 - f1-score (micro avg) 0.6564
2023-10-17 10:35:53,824 saving best model
2023-10-17 10:35:54,466 ----------------------------------------------------------------------------------------------------
2023-10-17 10:36:08,260 epoch 6 - iter 180/1809 - loss 0.02526723 - time (sec): 13.79 - samples/sec: 2751.63 - lr: 0.000016 - momentum: 0.000000
2023-10-17 10:36:21,717 epoch 6 - iter 360/1809 - loss 0.02266318 - time (sec): 27.25 - samples/sec: 2791.79 - lr: 0.000016 - momentum: 0.000000
2023-10-17 10:36:35,235 epoch 6 - iter 540/1809 - loss 0.02250374 - time (sec): 40.77 - samples/sec: 2768.29 - lr: 0.000016 - momentum: 0.000000
2023-10-17 10:36:48,648 epoch 6 - iter 720/1809 - loss 0.02114013 - time (sec): 54.18 - samples/sec: 2787.83 - lr: 0.000015 - momentum: 0.000000
2023-10-17 10:37:02,292 epoch 6 - iter 900/1809 - loss 0.02208737 - time (sec): 67.82 - samples/sec: 2787.14 - lr: 0.000015 - momentum: 0.000000
2023-10-17 10:37:15,201 epoch 6 - iter 1080/1809 - loss 0.02271038 - time (sec): 80.73 - samples/sec: 2800.43 - lr: 0.000015 - momentum: 0.000000
2023-10-17 10:37:29,400 epoch 6 - iter 1260/1809 - loss 0.02373796 - time (sec): 94.93 - samples/sec: 2783.97 - lr: 0.000014 - momentum: 0.000000
2023-10-17 10:37:43,390 epoch 6 - iter 1440/1809 - loss 0.02351275 - time (sec): 108.92 - samples/sec: 2769.13 - lr: 0.000014 - momentum: 0.000000
2023-10-17 10:37:56,318 epoch 6 - iter 1620/1809 - loss 0.02349186 - time (sec): 121.85 - samples/sec: 2791.97 - lr: 0.000014 - momentum: 0.000000
2023-10-17 10:38:09,258 epoch 6 - iter 1800/1809 - loss 0.02321458 - time (sec): 134.79 - samples/sec: 2801.94 - lr: 0.000013 - momentum: 0.000000
2023-10-17 10:38:09,960 ----------------------------------------------------------------------------------------------------
2023-10-17 10:38:09,961 EPOCH 6 done: loss 0.0231 - lr: 0.000013
2023-10-17 10:38:16,483 DEV : loss 0.317108690738678 - f1-score (micro avg) 0.6555
2023-10-17 10:38:16,533 ----------------------------------------------------------------------------------------------------
2023-10-17 10:38:30,845 epoch 7 - iter 180/1809 - loss 0.01666041 - time (sec): 14.31 - samples/sec: 2797.44 - lr: 0.000013 - momentum: 0.000000
2023-10-17 10:38:45,010 epoch 7 - iter 360/1809 - loss 0.01535944 - time (sec): 28.48 - samples/sec: 2763.71 - lr: 0.000013 - momentum: 0.000000
2023-10-17 10:38:59,190 epoch 7 - iter 540/1809 - loss 0.01585990 - time (sec): 42.66 - samples/sec: 2712.38 - lr: 0.000012 - momentum: 0.000000
2023-10-17 10:39:12,131 epoch 7 - iter 720/1809 - loss 0.01588240 - time (sec): 55.60 - samples/sec: 2743.87 - lr: 0.000012 - momentum: 0.000000
2023-10-17 10:39:24,799 epoch 7 - iter 900/1809 - loss 0.01567846 - time (sec): 68.26 - samples/sec: 2772.04 - lr: 0.000012 - momentum: 0.000000
2023-10-17 10:39:37,658 epoch 7 - iter 1080/1809 - loss 0.01573739 - time (sec): 81.12 - samples/sec: 2782.75 - lr: 0.000011 - momentum: 0.000000
2023-10-17 10:39:51,060 epoch 7 - iter 1260/1809 - loss 0.01492397 - time (sec): 94.53 - samples/sec: 2794.91 - lr: 0.000011 - momentum: 0.000000
2023-10-17 10:40:04,433 epoch 7 - iter 1440/1809 - loss 0.01512705 - time (sec): 107.90 - samples/sec: 2804.37 - lr: 0.000011 - momentum: 0.000000
2023-10-17 10:40:18,776 epoch 7 - iter 1620/1809 - loss 0.01537735 - time (sec): 122.24 - samples/sec: 2781.25 - lr: 0.000010 - momentum: 0.000000
2023-10-17 10:40:33,065 epoch 7 - iter 1800/1809 - loss 0.01577194 - time (sec): 136.53 - samples/sec: 2771.15 - lr: 0.000010 - momentum: 0.000000
2023-10-17 10:40:33,734 ----------------------------------------------------------------------------------------------------
2023-10-17 10:40:33,735 EPOCH 7 done: loss 0.0157 - lr: 0.000010
2023-10-17 10:40:40,913 DEV : loss 0.35636281967163086 - f1-score (micro avg) 0.6586
2023-10-17 10:40:40,989 saving best model
2023-10-17 10:40:41,627 ----------------------------------------------------------------------------------------------------
2023-10-17 10:40:55,327 epoch 8 - iter 180/1809 - loss 0.01233300 - time (sec): 13.70 - samples/sec: 2714.96 - lr: 0.000010 - momentum: 0.000000
2023-10-17 10:41:08,947 epoch 8 - iter 360/1809 - loss 0.01007325 - time (sec): 27.32 - samples/sec: 2768.24 - lr: 0.000009 - momentum: 0.000000
2023-10-17 10:41:22,905 epoch 8 - iter 540/1809 - loss 0.01079234 - time (sec): 41.28 - samples/sec: 2745.17 - lr: 0.000009 - momentum: 0.000000
2023-10-17 10:41:36,976 epoch 8 - iter 720/1809 - loss 0.01286652 - time (sec): 55.35 - samples/sec: 2726.46 - lr: 0.000009 - momentum: 0.000000
2023-10-17 10:41:51,431 epoch 8 - iter 900/1809 - loss 0.01195537 - time (sec): 69.80 - samples/sec: 2706.92 - lr: 0.000008 - momentum: 0.000000
2023-10-17 10:42:05,932 epoch 8 - iter 1080/1809 - loss 0.01176322 - time (sec): 84.30 - samples/sec: 2683.32 - lr: 0.000008 - momentum: 0.000000
2023-10-17 10:42:20,154 epoch 8 - iter 1260/1809 - loss 0.01134419 - time (sec): 98.52 - samples/sec: 2682.88 - lr: 0.000008 - momentum: 0.000000
2023-10-17 10:42:33,804 epoch 8 - iter 1440/1809 - loss 0.01129014 - time (sec): 112.18 - samples/sec: 2696.96 - lr: 0.000007 - momentum: 0.000000
2023-10-17 10:42:47,117 epoch 8 - iter 1620/1809 - loss 0.01089610 - time (sec): 125.49 - samples/sec: 2708.67 - lr: 0.000007 - momentum: 0.000000
2023-10-17 10:43:00,413 epoch 8 - iter 1800/1809 - loss 0.01091399 - time (sec): 138.78 - samples/sec: 2725.05 - lr: 0.000007 - momentum: 0.000000
2023-10-17 10:43:01,033 ----------------------------------------------------------------------------------------------------
2023-10-17 10:43:01,033 EPOCH 8 done: loss 0.0110 - lr: 0.000007
2023-10-17 10:43:07,612 DEV : loss 0.35512134432792664 - f1-score (micro avg) 0.6591
2023-10-17 10:43:07,659 saving best model
2023-10-17 10:43:08,266 ----------------------------------------------------------------------------------------------------
2023-10-17 10:43:21,490 epoch 9 - iter 180/1809 - loss 0.00903196 - time (sec): 13.22 - samples/sec: 2927.11 - lr: 0.000006 - momentum: 0.000000
2023-10-17 10:43:33,246 epoch 9 - iter 360/1809 - loss 0.00779450 - time (sec): 24.98 - samples/sec: 2999.25 - lr: 0.000006 - momentum: 0.000000
2023-10-17 10:43:45,896 epoch 9 - iter 540/1809 - loss 0.00679913 - time (sec): 37.63 - samples/sec: 2998.42 - lr: 0.000006 - momentum: 0.000000
2023-10-17 10:43:57,445 epoch 9 - iter 720/1809 - loss 0.00667016 - time (sec): 49.18 - samples/sec: 3080.76 - lr: 0.000005 - momentum: 0.000000
2023-10-17 10:44:09,143 epoch 9 - iter 900/1809 - loss 0.00738014 - time (sec): 60.88 - samples/sec: 3111.49 - lr: 0.000005 - momentum: 0.000000
2023-10-17 10:44:21,055 epoch 9 - iter 1080/1809 - loss 0.00679923 - time (sec): 72.79 - samples/sec: 3125.62 - lr: 0.000005 - momentum: 0.000000
2023-10-17 10:44:33,455 epoch 9 - iter 1260/1809 - loss 0.00660081 - time (sec): 85.19 - samples/sec: 3109.31 - lr: 0.000004 - momentum: 0.000000
2023-10-17 10:44:47,147 epoch 9 - iter 1440/1809 - loss 0.00659073 - time (sec): 98.88 - samples/sec: 3062.56 - lr: 0.000004 - momentum: 0.000000
2023-10-17 10:45:01,502 epoch 9 - iter 1620/1809 - loss 0.00679880 - time (sec): 113.23 - samples/sec: 3015.68 - lr: 0.000004 - momentum: 0.000000
2023-10-17 10:45:14,150 epoch 9 - iter 1800/1809 - loss 0.00670581 - time (sec): 125.88 - samples/sec: 3006.68 - lr: 0.000003 - momentum: 0.000000
2023-10-17 10:45:14,787 ----------------------------------------------------------------------------------------------------
2023-10-17 10:45:14,787 EPOCH 9 done: loss 0.0068 - lr: 0.000003
2023-10-17 10:45:21,264 DEV : loss 0.38741225004196167 - f1-score (micro avg) 0.6648
2023-10-17 10:45:21,310 saving best model
2023-10-17 10:45:21,908 ----------------------------------------------------------------------------------------------------
2023-10-17 10:45:35,707 epoch 10 - iter 180/1809 - loss 0.00757151 - time (sec): 13.80 - samples/sec: 2784.91 - lr: 0.000003 - momentum: 0.000000
2023-10-17 10:45:49,725 epoch 10 - iter 360/1809 - loss 0.00689107 - time (sec): 27.82 - samples/sec: 2720.52 - lr: 0.000003 - momentum: 0.000000
2023-10-17 10:46:04,659 epoch 10 - iter 540/1809 - loss 0.00613861 - time (sec): 42.75 - samples/sec: 2647.42 - lr: 0.000002 - momentum: 0.000000
2023-10-17 10:46:18,293 epoch 10 - iter 720/1809 - loss 0.00588222 - time (sec): 56.38 - samples/sec: 2663.59 - lr: 0.000002 - momentum: 0.000000
2023-10-17 10:46:32,068 epoch 10 - iter 900/1809 - loss 0.00544332 - time (sec): 70.16 - samples/sec: 2658.14 - lr: 0.000002 - momentum: 0.000000
2023-10-17 10:46:45,517 epoch 10 - iter 1080/1809 - loss 0.00530158 - time (sec): 83.61 - samples/sec: 2688.74 - lr: 0.000001 - momentum: 0.000000
2023-10-17 10:46:59,225 epoch 10 - iter 1260/1809 - loss 0.00497582 - time (sec): 97.32 - samples/sec: 2719.02 - lr: 0.000001 - momentum: 0.000000
2023-10-17 10:47:12,932 epoch 10 - iter 1440/1809 - loss 0.00553063 - time (sec): 111.02 - samples/sec: 2716.47 - lr: 0.000001 - momentum: 0.000000
2023-10-17 10:47:27,001 epoch 10 - iter 1620/1809 - loss 0.00533905 - time (sec): 125.09 - samples/sec: 2716.76 - lr: 0.000000 - momentum: 0.000000
2023-10-17 10:47:41,204 epoch 10 - iter 1800/1809 - loss 0.00529427 - time (sec): 139.29 - samples/sec: 2715.81 - lr: 0.000000 - momentum: 0.000000
2023-10-17 10:47:41,818 ----------------------------------------------------------------------------------------------------
2023-10-17 10:47:41,818 EPOCH 10 done: loss 0.0054 - lr: 0.000000
2023-10-17 10:47:48,079 DEV : loss 0.40038371086120605 - f1-score (micro avg) 0.6651
2023-10-17 10:47:48,122 saving best model
2023-10-17 10:47:49,288 ----------------------------------------------------------------------------------------------------
2023-10-17 10:47:49,290 Loading model from best epoch ...
2023-10-17 10:47:51,019 SequenceTagger predicts: Dictionary with 13 tags: O, S-loc, B-loc, E-loc, I-loc, S-pers, B-pers, E-pers, I-pers, S-org, B-org, E-org, I-org
2023-10-17 10:48:00,015
Results:
- F-score (micro) 0.6719
- F-score (macro) 0.5432
- Accuracy 0.5187
By class:
precision recall f1-score support
loc 0.6531 0.8156 0.7254 591
pers 0.5876 0.7423 0.6559 357
org 0.2931 0.2152 0.2482 79
micro avg 0.6127 0.7439 0.6719 1027
macro avg 0.5113 0.5910 0.5432 1027
weighted avg 0.6026 0.7439 0.6645 1027
2023-10-17 10:48:00,015 ----------------------------------------------------------------------------------------------------
|