File size: 25,535 Bytes
19acd5b |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 |
2023-10-11 11:23:41,250 ----------------------------------------------------------------------------------------------------
2023-10-11 11:23:41,253 Model: "SequenceTagger(
(embeddings): ByT5Embeddings(
(model): T5EncoderModel(
(shared): Embedding(384, 1472)
(encoder): T5Stack(
(embed_tokens): Embedding(384, 1472)
(block): ModuleList(
(0): T5Block(
(layer): ModuleList(
(0): T5LayerSelfAttention(
(SelfAttention): T5Attention(
(q): Linear(in_features=1472, out_features=384, bias=False)
(k): Linear(in_features=1472, out_features=384, bias=False)
(v): Linear(in_features=1472, out_features=384, bias=False)
(o): Linear(in_features=384, out_features=1472, bias=False)
(relative_attention_bias): Embedding(32, 6)
)
(layer_norm): FusedRMSNorm(torch.Size([1472]), eps=1e-06, elementwise_affine=True)
(dropout): Dropout(p=0.1, inplace=False)
)
(1): T5LayerFF(
(DenseReluDense): T5DenseGatedActDense(
(wi_0): Linear(in_features=1472, out_features=3584, bias=False)
(wi_1): Linear(in_features=1472, out_features=3584, bias=False)
(wo): Linear(in_features=3584, out_features=1472, bias=False)
(dropout): Dropout(p=0.1, inplace=False)
(act): NewGELUActivation()
)
(layer_norm): FusedRMSNorm(torch.Size([1472]), eps=1e-06, elementwise_affine=True)
(dropout): Dropout(p=0.1, inplace=False)
)
)
)
(1-11): 11 x T5Block(
(layer): ModuleList(
(0): T5LayerSelfAttention(
(SelfAttention): T5Attention(
(q): Linear(in_features=1472, out_features=384, bias=False)
(k): Linear(in_features=1472, out_features=384, bias=False)
(v): Linear(in_features=1472, out_features=384, bias=False)
(o): Linear(in_features=384, out_features=1472, bias=False)
)
(layer_norm): FusedRMSNorm(torch.Size([1472]), eps=1e-06, elementwise_affine=True)
(dropout): Dropout(p=0.1, inplace=False)
)
(1): T5LayerFF(
(DenseReluDense): T5DenseGatedActDense(
(wi_0): Linear(in_features=1472, out_features=3584, bias=False)
(wi_1): Linear(in_features=1472, out_features=3584, bias=False)
(wo): Linear(in_features=3584, out_features=1472, bias=False)
(dropout): Dropout(p=0.1, inplace=False)
(act): NewGELUActivation()
)
(layer_norm): FusedRMSNorm(torch.Size([1472]), eps=1e-06, elementwise_affine=True)
(dropout): Dropout(p=0.1, inplace=False)
)
)
)
)
(final_layer_norm): FusedRMSNorm(torch.Size([1472]), eps=1e-06, elementwise_affine=True)
(dropout): Dropout(p=0.1, inplace=False)
)
)
)
(locked_dropout): LockedDropout(p=0.5)
(linear): Linear(in_features=1472, out_features=17, bias=True)
(loss_function): CrossEntropyLoss()
)"
2023-10-11 11:23:41,253 ----------------------------------------------------------------------------------------------------
2023-10-11 11:23:41,253 MultiCorpus: 7142 train + 698 dev + 2570 test sentences
- NER_HIPE_2022 Corpus: 7142 train + 698 dev + 2570 test sentences - /root/.flair/datasets/ner_hipe_2022/v2.1/newseye/fr/with_doc_seperator
2023-10-11 11:23:41,253 ----------------------------------------------------------------------------------------------------
2023-10-11 11:23:41,253 Train: 7142 sentences
2023-10-11 11:23:41,253 (train_with_dev=False, train_with_test=False)
2023-10-11 11:23:41,253 ----------------------------------------------------------------------------------------------------
2023-10-11 11:23:41,253 Training Params:
2023-10-11 11:23:41,253 - learning_rate: "0.00015"
2023-10-11 11:23:41,253 - mini_batch_size: "4"
2023-10-11 11:23:41,254 - max_epochs: "10"
2023-10-11 11:23:41,254 - shuffle: "True"
2023-10-11 11:23:41,254 ----------------------------------------------------------------------------------------------------
2023-10-11 11:23:41,254 Plugins:
2023-10-11 11:23:41,254 - TensorboardLogger
2023-10-11 11:23:41,254 - LinearScheduler | warmup_fraction: '0.1'
2023-10-11 11:23:41,254 ----------------------------------------------------------------------------------------------------
2023-10-11 11:23:41,254 Final evaluation on model from best epoch (best-model.pt)
2023-10-11 11:23:41,254 - metric: "('micro avg', 'f1-score')"
2023-10-11 11:23:41,254 ----------------------------------------------------------------------------------------------------
2023-10-11 11:23:41,254 Computation:
2023-10-11 11:23:41,254 - compute on device: cuda:0
2023-10-11 11:23:41,254 - embedding storage: none
2023-10-11 11:23:41,254 ----------------------------------------------------------------------------------------------------
2023-10-11 11:23:41,254 Model training base path: "hmbench-newseye/fr-hmbyt5-preliminary/byt5-small-historic-multilingual-span20-flax-bs4-wsFalse-e10-lr0.00015-poolingfirst-layers-1-crfFalse-3"
2023-10-11 11:23:41,255 ----------------------------------------------------------------------------------------------------
2023-10-11 11:23:41,255 ----------------------------------------------------------------------------------------------------
2023-10-11 11:23:41,255 Logging anything other than scalars to TensorBoard is currently not supported.
2023-10-11 11:24:36,735 epoch 1 - iter 178/1786 - loss 2.81334662 - time (sec): 55.48 - samples/sec: 485.36 - lr: 0.000015 - momentum: 0.000000
2023-10-11 11:25:30,124 epoch 1 - iter 356/1786 - loss 2.65360553 - time (sec): 108.87 - samples/sec: 465.62 - lr: 0.000030 - momentum: 0.000000
2023-10-11 11:26:25,121 epoch 1 - iter 534/1786 - loss 2.38474398 - time (sec): 163.86 - samples/sec: 454.74 - lr: 0.000045 - momentum: 0.000000
2023-10-11 11:27:20,005 epoch 1 - iter 712/1786 - loss 2.09942018 - time (sec): 218.75 - samples/sec: 451.36 - lr: 0.000060 - momentum: 0.000000
2023-10-11 11:28:22,609 epoch 1 - iter 890/1786 - loss 1.82583585 - time (sec): 281.35 - samples/sec: 443.83 - lr: 0.000075 - momentum: 0.000000
2023-10-11 11:29:19,386 epoch 1 - iter 1068/1786 - loss 1.62155219 - time (sec): 338.13 - samples/sec: 439.95 - lr: 0.000090 - momentum: 0.000000
2023-10-11 11:30:14,335 epoch 1 - iter 1246/1786 - loss 1.44125903 - time (sec): 393.08 - samples/sec: 443.60 - lr: 0.000105 - momentum: 0.000000
2023-10-11 11:31:09,952 epoch 1 - iter 1424/1786 - loss 1.30760460 - time (sec): 448.70 - samples/sec: 443.19 - lr: 0.000120 - momentum: 0.000000
2023-10-11 11:32:05,906 epoch 1 - iter 1602/1786 - loss 1.19189571 - time (sec): 504.65 - samples/sec: 443.96 - lr: 0.000134 - momentum: 0.000000
2023-10-11 11:33:03,148 epoch 1 - iter 1780/1786 - loss 1.10257673 - time (sec): 561.89 - samples/sec: 441.25 - lr: 0.000149 - momentum: 0.000000
2023-10-11 11:33:04,915 ----------------------------------------------------------------------------------------------------
2023-10-11 11:33:04,916 EPOCH 1 done: loss 1.0997 - lr: 0.000149
2023-10-11 11:33:25,813 DEV : loss 0.17743352055549622 - f1-score (micro avg) 0.6248
2023-10-11 11:33:25,849 saving best model
2023-10-11 11:33:26,769 ----------------------------------------------------------------------------------------------------
2023-10-11 11:34:26,264 epoch 2 - iter 178/1786 - loss 0.16761718 - time (sec): 59.49 - samples/sec: 441.01 - lr: 0.000148 - momentum: 0.000000
2023-10-11 11:35:22,903 epoch 2 - iter 356/1786 - loss 0.17182138 - time (sec): 116.13 - samples/sec: 443.29 - lr: 0.000147 - momentum: 0.000000
2023-10-11 11:36:17,446 epoch 2 - iter 534/1786 - loss 0.16152548 - time (sec): 170.67 - samples/sec: 441.52 - lr: 0.000145 - momentum: 0.000000
2023-10-11 11:37:12,824 epoch 2 - iter 712/1786 - loss 0.14888668 - time (sec): 226.05 - samples/sec: 445.71 - lr: 0.000143 - momentum: 0.000000
2023-10-11 11:38:04,419 epoch 2 - iter 890/1786 - loss 0.14407949 - time (sec): 277.65 - samples/sec: 448.14 - lr: 0.000142 - momentum: 0.000000
2023-10-11 11:38:58,004 epoch 2 - iter 1068/1786 - loss 0.14005413 - time (sec): 331.23 - samples/sec: 453.55 - lr: 0.000140 - momentum: 0.000000
2023-10-11 11:39:50,454 epoch 2 - iter 1246/1786 - loss 0.13811135 - time (sec): 383.68 - samples/sec: 455.24 - lr: 0.000138 - momentum: 0.000000
2023-10-11 11:40:45,636 epoch 2 - iter 1424/1786 - loss 0.13523123 - time (sec): 438.86 - samples/sec: 451.08 - lr: 0.000137 - momentum: 0.000000
2023-10-11 11:41:43,985 epoch 2 - iter 1602/1786 - loss 0.13268221 - time (sec): 497.21 - samples/sec: 447.57 - lr: 0.000135 - momentum: 0.000000
2023-10-11 11:42:37,848 epoch 2 - iter 1780/1786 - loss 0.12982701 - time (sec): 551.08 - samples/sec: 449.89 - lr: 0.000133 - momentum: 0.000000
2023-10-11 11:42:39,599 ----------------------------------------------------------------------------------------------------
2023-10-11 11:42:39,599 EPOCH 2 done: loss 0.1296 - lr: 0.000133
2023-10-11 11:43:01,563 DEV : loss 0.1018710657954216 - f1-score (micro avg) 0.7677
2023-10-11 11:43:01,593 saving best model
2023-10-11 11:43:04,216 ----------------------------------------------------------------------------------------------------
2023-10-11 11:43:55,588 epoch 3 - iter 178/1786 - loss 0.06089358 - time (sec): 51.37 - samples/sec: 464.22 - lr: 0.000132 - momentum: 0.000000
2023-10-11 11:44:47,335 epoch 3 - iter 356/1786 - loss 0.06083272 - time (sec): 103.11 - samples/sec: 474.26 - lr: 0.000130 - momentum: 0.000000
2023-10-11 11:45:39,079 epoch 3 - iter 534/1786 - loss 0.06301108 - time (sec): 154.86 - samples/sec: 472.60 - lr: 0.000128 - momentum: 0.000000
2023-10-11 11:46:31,108 epoch 3 - iter 712/1786 - loss 0.06615584 - time (sec): 206.89 - samples/sec: 474.57 - lr: 0.000127 - momentum: 0.000000
2023-10-11 11:47:24,530 epoch 3 - iter 890/1786 - loss 0.07031467 - time (sec): 260.31 - samples/sec: 472.06 - lr: 0.000125 - momentum: 0.000000
2023-10-11 11:48:18,485 epoch 3 - iter 1068/1786 - loss 0.07283988 - time (sec): 314.26 - samples/sec: 469.38 - lr: 0.000123 - momentum: 0.000000
2023-10-11 11:49:17,279 epoch 3 - iter 1246/1786 - loss 0.07528099 - time (sec): 373.06 - samples/sec: 466.84 - lr: 0.000122 - momentum: 0.000000
2023-10-11 11:50:13,121 epoch 3 - iter 1424/1786 - loss 0.07404515 - time (sec): 428.90 - samples/sec: 461.84 - lr: 0.000120 - momentum: 0.000000
2023-10-11 11:51:06,569 epoch 3 - iter 1602/1786 - loss 0.07276736 - time (sec): 482.35 - samples/sec: 462.80 - lr: 0.000118 - momentum: 0.000000
2023-10-11 11:51:59,561 epoch 3 - iter 1780/1786 - loss 0.07349545 - time (sec): 535.34 - samples/sec: 463.53 - lr: 0.000117 - momentum: 0.000000
2023-10-11 11:52:01,090 ----------------------------------------------------------------------------------------------------
2023-10-11 11:52:01,091 EPOCH 3 done: loss 0.0738 - lr: 0.000117
2023-10-11 11:52:23,243 DEV : loss 0.11691577732563019 - f1-score (micro avg) 0.7871
2023-10-11 11:52:23,273 saving best model
2023-10-11 11:52:25,840 ----------------------------------------------------------------------------------------------------
2023-10-11 11:53:18,856 epoch 4 - iter 178/1786 - loss 0.03874746 - time (sec): 53.01 - samples/sec: 453.23 - lr: 0.000115 - momentum: 0.000000
2023-10-11 11:54:17,169 epoch 4 - iter 356/1786 - loss 0.04840883 - time (sec): 111.32 - samples/sec: 440.32 - lr: 0.000113 - momentum: 0.000000
2023-10-11 11:55:15,656 epoch 4 - iter 534/1786 - loss 0.04781593 - time (sec): 169.81 - samples/sec: 445.25 - lr: 0.000112 - momentum: 0.000000
2023-10-11 11:56:12,421 epoch 4 - iter 712/1786 - loss 0.05099290 - time (sec): 226.58 - samples/sec: 442.20 - lr: 0.000110 - momentum: 0.000000
2023-10-11 11:57:11,612 epoch 4 - iter 890/1786 - loss 0.05175638 - time (sec): 285.77 - samples/sec: 440.13 - lr: 0.000108 - momentum: 0.000000
2023-10-11 11:58:05,630 epoch 4 - iter 1068/1786 - loss 0.05367502 - time (sec): 339.79 - samples/sec: 437.42 - lr: 0.000107 - momentum: 0.000000
2023-10-11 11:59:01,799 epoch 4 - iter 1246/1786 - loss 0.05394290 - time (sec): 395.95 - samples/sec: 438.77 - lr: 0.000105 - momentum: 0.000000
2023-10-11 11:59:56,173 epoch 4 - iter 1424/1786 - loss 0.05369425 - time (sec): 450.33 - samples/sec: 439.57 - lr: 0.000103 - momentum: 0.000000
2023-10-11 12:00:51,253 epoch 4 - iter 1602/1786 - loss 0.05297193 - time (sec): 505.41 - samples/sec: 441.40 - lr: 0.000102 - momentum: 0.000000
2023-10-11 12:01:47,664 epoch 4 - iter 1780/1786 - loss 0.05177312 - time (sec): 561.82 - samples/sec: 441.43 - lr: 0.000100 - momentum: 0.000000
2023-10-11 12:01:49,317 ----------------------------------------------------------------------------------------------------
2023-10-11 12:01:49,317 EPOCH 4 done: loss 0.0517 - lr: 0.000100
2023-10-11 12:02:11,096 DEV : loss 0.1411616951227188 - f1-score (micro avg) 0.7951
2023-10-11 12:02:11,133 saving best model
2023-10-11 12:02:13,751 ----------------------------------------------------------------------------------------------------
2023-10-11 12:03:07,905 epoch 5 - iter 178/1786 - loss 0.03072434 - time (sec): 54.15 - samples/sec: 463.84 - lr: 0.000098 - momentum: 0.000000
2023-10-11 12:04:04,994 epoch 5 - iter 356/1786 - loss 0.03893563 - time (sec): 111.24 - samples/sec: 454.86 - lr: 0.000097 - momentum: 0.000000
2023-10-11 12:04:59,333 epoch 5 - iter 534/1786 - loss 0.03550559 - time (sec): 165.58 - samples/sec: 456.77 - lr: 0.000095 - momentum: 0.000000
2023-10-11 12:05:54,205 epoch 5 - iter 712/1786 - loss 0.03476839 - time (sec): 220.45 - samples/sec: 452.42 - lr: 0.000093 - momentum: 0.000000
2023-10-11 12:06:47,739 epoch 5 - iter 890/1786 - loss 0.03433559 - time (sec): 273.98 - samples/sec: 451.14 - lr: 0.000092 - momentum: 0.000000
2023-10-11 12:07:41,693 epoch 5 - iter 1068/1786 - loss 0.03434337 - time (sec): 327.94 - samples/sec: 450.14 - lr: 0.000090 - momentum: 0.000000
2023-10-11 12:08:37,890 epoch 5 - iter 1246/1786 - loss 0.03476657 - time (sec): 384.13 - samples/sec: 450.22 - lr: 0.000088 - momentum: 0.000000
2023-10-11 12:09:36,033 epoch 5 - iter 1424/1786 - loss 0.03524109 - time (sec): 442.28 - samples/sec: 446.25 - lr: 0.000087 - momentum: 0.000000
2023-10-11 12:10:33,305 epoch 5 - iter 1602/1786 - loss 0.03503721 - time (sec): 499.55 - samples/sec: 445.01 - lr: 0.000085 - momentum: 0.000000
2023-10-11 12:11:30,163 epoch 5 - iter 1780/1786 - loss 0.03719160 - time (sec): 556.41 - samples/sec: 445.83 - lr: 0.000083 - momentum: 0.000000
2023-10-11 12:11:31,834 ----------------------------------------------------------------------------------------------------
2023-10-11 12:11:31,835 EPOCH 5 done: loss 0.0373 - lr: 0.000083
2023-10-11 12:11:54,146 DEV : loss 0.1446281224489212 - f1-score (micro avg) 0.7989
2023-10-11 12:11:54,177 saving best model
2023-10-11 12:11:56,879 ----------------------------------------------------------------------------------------------------
2023-10-11 12:12:51,045 epoch 6 - iter 178/1786 - loss 0.03460331 - time (sec): 54.16 - samples/sec: 480.26 - lr: 0.000082 - momentum: 0.000000
2023-10-11 12:13:45,100 epoch 6 - iter 356/1786 - loss 0.03107979 - time (sec): 108.22 - samples/sec: 458.69 - lr: 0.000080 - momentum: 0.000000
2023-10-11 12:14:39,394 epoch 6 - iter 534/1786 - loss 0.03181132 - time (sec): 162.51 - samples/sec: 453.26 - lr: 0.000078 - momentum: 0.000000
2023-10-11 12:15:33,225 epoch 6 - iter 712/1786 - loss 0.03080483 - time (sec): 216.34 - samples/sec: 458.64 - lr: 0.000077 - momentum: 0.000000
2023-10-11 12:16:27,715 epoch 6 - iter 890/1786 - loss 0.02989970 - time (sec): 270.83 - samples/sec: 456.86 - lr: 0.000075 - momentum: 0.000000
2023-10-11 12:17:20,578 epoch 6 - iter 1068/1786 - loss 0.02943783 - time (sec): 323.70 - samples/sec: 454.31 - lr: 0.000073 - momentum: 0.000000
2023-10-11 12:18:16,492 epoch 6 - iter 1246/1786 - loss 0.02814408 - time (sec): 379.61 - samples/sec: 452.62 - lr: 0.000072 - momentum: 0.000000
2023-10-11 12:19:11,158 epoch 6 - iter 1424/1786 - loss 0.02929556 - time (sec): 434.28 - samples/sec: 456.28 - lr: 0.000070 - momentum: 0.000000
2023-10-11 12:20:06,940 epoch 6 - iter 1602/1786 - loss 0.02877508 - time (sec): 490.06 - samples/sec: 455.44 - lr: 0.000068 - momentum: 0.000000
2023-10-11 12:21:07,155 epoch 6 - iter 1780/1786 - loss 0.02880009 - time (sec): 550.27 - samples/sec: 450.81 - lr: 0.000067 - momentum: 0.000000
2023-10-11 12:21:08,998 ----------------------------------------------------------------------------------------------------
2023-10-11 12:21:08,998 EPOCH 6 done: loss 0.0287 - lr: 0.000067
2023-10-11 12:21:31,729 DEV : loss 0.18198005855083466 - f1-score (micro avg) 0.7913
2023-10-11 12:21:31,760 ----------------------------------------------------------------------------------------------------
2023-10-11 12:22:27,207 epoch 7 - iter 178/1786 - loss 0.02502610 - time (sec): 55.45 - samples/sec: 442.13 - lr: 0.000065 - momentum: 0.000000
2023-10-11 12:23:18,505 epoch 7 - iter 356/1786 - loss 0.02577891 - time (sec): 106.74 - samples/sec: 447.65 - lr: 0.000063 - momentum: 0.000000
2023-10-11 12:24:11,729 epoch 7 - iter 534/1786 - loss 0.02380424 - time (sec): 159.97 - samples/sec: 456.73 - lr: 0.000062 - momentum: 0.000000
2023-10-11 12:25:03,073 epoch 7 - iter 712/1786 - loss 0.02361198 - time (sec): 211.31 - samples/sec: 459.78 - lr: 0.000060 - momentum: 0.000000
2023-10-11 12:25:54,713 epoch 7 - iter 890/1786 - loss 0.02408280 - time (sec): 262.95 - samples/sec: 465.29 - lr: 0.000058 - momentum: 0.000000
2023-10-11 12:26:47,777 epoch 7 - iter 1068/1786 - loss 0.02326326 - time (sec): 316.01 - samples/sec: 468.21 - lr: 0.000057 - momentum: 0.000000
2023-10-11 12:27:42,548 epoch 7 - iter 1246/1786 - loss 0.02207530 - time (sec): 370.79 - samples/sec: 466.64 - lr: 0.000055 - momentum: 0.000000
2023-10-11 12:28:34,067 epoch 7 - iter 1424/1786 - loss 0.02234953 - time (sec): 422.31 - samples/sec: 469.14 - lr: 0.000053 - momentum: 0.000000
2023-10-11 12:29:25,675 epoch 7 - iter 1602/1786 - loss 0.02289223 - time (sec): 473.91 - samples/sec: 471.10 - lr: 0.000052 - momentum: 0.000000
2023-10-11 12:30:17,009 epoch 7 - iter 1780/1786 - loss 0.02244176 - time (sec): 525.25 - samples/sec: 472.48 - lr: 0.000050 - momentum: 0.000000
2023-10-11 12:30:18,464 ----------------------------------------------------------------------------------------------------
2023-10-11 12:30:18,465 EPOCH 7 done: loss 0.0224 - lr: 0.000050
2023-10-11 12:30:38,760 DEV : loss 0.19651886820793152 - f1-score (micro avg) 0.7944
2023-10-11 12:30:38,790 ----------------------------------------------------------------------------------------------------
2023-10-11 12:31:30,271 epoch 8 - iter 178/1786 - loss 0.01170122 - time (sec): 51.48 - samples/sec: 479.72 - lr: 0.000048 - momentum: 0.000000
2023-10-11 12:32:21,756 epoch 8 - iter 356/1786 - loss 0.01261849 - time (sec): 102.96 - samples/sec: 480.14 - lr: 0.000047 - momentum: 0.000000
2023-10-11 12:33:11,986 epoch 8 - iter 534/1786 - loss 0.01077078 - time (sec): 153.19 - samples/sec: 474.98 - lr: 0.000045 - momentum: 0.000000
2023-10-11 12:34:02,991 epoch 8 - iter 712/1786 - loss 0.01119283 - time (sec): 204.20 - samples/sec: 471.16 - lr: 0.000043 - momentum: 0.000000
2023-10-11 12:34:54,856 epoch 8 - iter 890/1786 - loss 0.01302730 - time (sec): 256.06 - samples/sec: 469.17 - lr: 0.000042 - momentum: 0.000000
2023-10-11 12:35:52,337 epoch 8 - iter 1068/1786 - loss 0.01505220 - time (sec): 313.55 - samples/sec: 467.37 - lr: 0.000040 - momentum: 0.000000
2023-10-11 12:36:46,291 epoch 8 - iter 1246/1786 - loss 0.01529191 - time (sec): 367.50 - samples/sec: 469.17 - lr: 0.000038 - momentum: 0.000000
2023-10-11 12:37:40,067 epoch 8 - iter 1424/1786 - loss 0.01565327 - time (sec): 421.28 - samples/sec: 472.27 - lr: 0.000037 - momentum: 0.000000
2023-10-11 12:38:33,123 epoch 8 - iter 1602/1786 - loss 0.01640331 - time (sec): 474.33 - samples/sec: 473.18 - lr: 0.000035 - momentum: 0.000000
2023-10-11 12:39:24,506 epoch 8 - iter 1780/1786 - loss 0.01586581 - time (sec): 525.71 - samples/sec: 471.92 - lr: 0.000033 - momentum: 0.000000
2023-10-11 12:39:26,070 ----------------------------------------------------------------------------------------------------
2023-10-11 12:39:26,070 EPOCH 8 done: loss 0.0158 - lr: 0.000033
2023-10-11 12:39:47,196 DEV : loss 0.20342709124088287 - f1-score (micro avg) 0.8005
2023-10-11 12:39:47,225 saving best model
2023-10-11 12:39:49,797 ----------------------------------------------------------------------------------------------------
2023-10-11 12:40:41,572 epoch 9 - iter 178/1786 - loss 0.01473686 - time (sec): 51.77 - samples/sec: 460.70 - lr: 0.000032 - momentum: 0.000000
2023-10-11 12:41:32,765 epoch 9 - iter 356/1786 - loss 0.01118027 - time (sec): 102.96 - samples/sec: 453.89 - lr: 0.000030 - momentum: 0.000000
2023-10-11 12:42:23,635 epoch 9 - iter 534/1786 - loss 0.01245056 - time (sec): 153.83 - samples/sec: 447.25 - lr: 0.000028 - momentum: 0.000000
2023-10-11 12:43:17,012 epoch 9 - iter 712/1786 - loss 0.01164884 - time (sec): 207.21 - samples/sec: 457.99 - lr: 0.000027 - momentum: 0.000000
2023-10-11 12:44:09,593 epoch 9 - iter 890/1786 - loss 0.01172585 - time (sec): 259.79 - samples/sec: 462.94 - lr: 0.000025 - momentum: 0.000000
2023-10-11 12:45:03,569 epoch 9 - iter 1068/1786 - loss 0.01162303 - time (sec): 313.77 - samples/sec: 465.80 - lr: 0.000023 - momentum: 0.000000
2023-10-11 12:45:58,302 epoch 9 - iter 1246/1786 - loss 0.01213149 - time (sec): 368.50 - samples/sec: 468.41 - lr: 0.000022 - momentum: 0.000000
2023-10-11 12:46:54,840 epoch 9 - iter 1424/1786 - loss 0.01176088 - time (sec): 425.04 - samples/sec: 467.35 - lr: 0.000020 - momentum: 0.000000
2023-10-11 12:47:47,993 epoch 9 - iter 1602/1786 - loss 0.01190777 - time (sec): 478.19 - samples/sec: 467.50 - lr: 0.000018 - momentum: 0.000000
2023-10-11 12:48:41,133 epoch 9 - iter 1780/1786 - loss 0.01190889 - time (sec): 531.33 - samples/sec: 466.14 - lr: 0.000017 - momentum: 0.000000
2023-10-11 12:48:42,969 ----------------------------------------------------------------------------------------------------
2023-10-11 12:48:42,969 EPOCH 9 done: loss 0.0120 - lr: 0.000017
2023-10-11 12:49:04,098 DEV : loss 0.21383821964263916 - f1-score (micro avg) 0.7934
2023-10-11 12:49:04,137 ----------------------------------------------------------------------------------------------------
2023-10-11 12:49:59,581 epoch 10 - iter 178/1786 - loss 0.00900177 - time (sec): 55.44 - samples/sec: 450.00 - lr: 0.000015 - momentum: 0.000000
2023-10-11 12:50:51,193 epoch 10 - iter 356/1786 - loss 0.01022974 - time (sec): 107.05 - samples/sec: 449.40 - lr: 0.000013 - momentum: 0.000000
2023-10-11 12:51:44,986 epoch 10 - iter 534/1786 - loss 0.00925737 - time (sec): 160.85 - samples/sec: 452.35 - lr: 0.000012 - momentum: 0.000000
2023-10-11 12:52:43,086 epoch 10 - iter 712/1786 - loss 0.00868107 - time (sec): 218.95 - samples/sec: 449.05 - lr: 0.000010 - momentum: 0.000000
2023-10-11 12:53:39,352 epoch 10 - iter 890/1786 - loss 0.00865781 - time (sec): 275.21 - samples/sec: 451.32 - lr: 0.000008 - momentum: 0.000000
2023-10-11 12:54:34,765 epoch 10 - iter 1068/1786 - loss 0.00851102 - time (sec): 330.62 - samples/sec: 449.74 - lr: 0.000007 - momentum: 0.000000
2023-10-11 12:55:30,317 epoch 10 - iter 1246/1786 - loss 0.00848258 - time (sec): 386.18 - samples/sec: 447.18 - lr: 0.000005 - momentum: 0.000000
2023-10-11 12:56:26,280 epoch 10 - iter 1424/1786 - loss 0.00850758 - time (sec): 442.14 - samples/sec: 446.92 - lr: 0.000003 - momentum: 0.000000
2023-10-11 12:57:23,983 epoch 10 - iter 1602/1786 - loss 0.00827861 - time (sec): 499.84 - samples/sec: 445.08 - lr: 0.000002 - momentum: 0.000000
2023-10-11 12:58:21,558 epoch 10 - iter 1780/1786 - loss 0.00830503 - time (sec): 557.42 - samples/sec: 445.35 - lr: 0.000000 - momentum: 0.000000
2023-10-11 12:58:23,007 ----------------------------------------------------------------------------------------------------
2023-10-11 12:58:23,007 EPOCH 10 done: loss 0.0083 - lr: 0.000000
2023-10-11 12:58:45,352 DEV : loss 0.2220211774110794 - f1-score (micro avg) 0.7952
2023-10-11 12:58:46,381 ----------------------------------------------------------------------------------------------------
2023-10-11 12:58:46,383 Loading model from best epoch ...
2023-10-11 12:58:50,444 SequenceTagger predicts: Dictionary with 17 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, S-HumanProd, B-HumanProd, E-HumanProd, I-HumanProd
2023-10-11 13:00:06,150
Results:
- F-score (micro) 0.7251
- F-score (macro) 0.6453
- Accuracy 0.5829
By class:
precision recall f1-score support
LOC 0.7389 0.7443 0.7416 1095
PER 0.7830 0.7846 0.7838 1012
ORG 0.5083 0.5994 0.5501 357
HumanProd 0.4074 0.6667 0.5057 33
micro avg 0.7118 0.7389 0.7251 2497
macro avg 0.6094 0.6987 0.6453 2497
weighted avg 0.7194 0.7389 0.7282 2497
2023-10-11 13:00:06,150 ----------------------------------------------------------------------------------------------------
|