Skip to content
Switch branches/tags

Name already in use

A tag already exists with the provided branch name. Many Git commands accept both tag and branch names, so creating this branch may cause unexpected behavior. Are you sure you want to create this branch?
Go to file
Cannot retrieve contributors at this time
INFO:gluonnlp:06:07:39 Namespace(accumulate=6, batch_size=4, bert_dataset='book_corpus_wiki_en_uncased', bert_model='bert_24_1024_16', debug=False, doc_stride=128, epochs=2, gpu=0, load_feature_from_pickle=False, log_interval=50, lr=3e-05, max_answer_length=30, max_query_length=64, max_seq_length=384, model_parameters=None, n_best_size=20, null_score_diff_threshold=0.0, only_predict=False, optimizer='adam', output_dir='./output_dir', pretrained_bert_parameters=None, sentencepiece=None, test_batch_size=24, training_steps=None, uncased=True, version_2=False, warmup_ratio=0.1)
INFO:gluonnlp:06:07:39 Using gradient accumulation. Effective batch size = 24
INFO:gluonnlp:06:08:35 Loading train data...
INFO:gluonnlp:06:08:36 Number of records in Train data:87599
INFO:gluonnlp:06:09:19 The number of examples after preprocessing:88641
INFO:gluonnlp:06:09:19 Start Training
INFO:gluonnlp:06:10:43 Epoch: 0, Batch: 299/22161, Loss=5.4315, lr=0.0000020 Time cost=83.7 Thoughput=14.33 samples/s
INFO:gluonnlp:06:12:07 Epoch: 0, Batch: 599/22161, Loss=5.1601, lr=0.0000041 Time cost=83.8 Thoughput=14.32 samples/s
INFO:gluonnlp:06:13:31 Epoch: 0, Batch: 899/22161, Loss=4.5888, lr=0.0000061 Time cost=84.0 Thoughput=14.28 samples/s
INFO:gluonnlp:06:14:55 Epoch: 0, Batch: 1199/22161, Loss=3.8660, lr=0.0000081 Time cost=84.4 Thoughput=14.22 samples/s
INFO:gluonnlp:06:16:20 Epoch: 0, Batch: 1499/22161, Loss=2.6834, lr=0.0000102 Time cost=84.8 Thoughput=14.16 samples/s
INFO:gluonnlp:06:17:44 Epoch: 0, Batch: 1799/22161, Loss=2.1346, lr=0.0000122 Time cost=84.1 Thoughput=14.27 samples/s
INFO:gluonnlp:06:19:06 Epoch: 0, Batch: 2099/22161, Loss=1.7814, lr=0.0000142 Time cost=82.3 Thoughput=14.59 samples/s
INFO:gluonnlp:06:20:29 Epoch: 0, Batch: 2399/22161, Loss=1.6098, lr=0.0000163 Time cost=82.5 Thoughput=14.54 samples/s
INFO:gluonnlp:06:21:51 Epoch: 0, Batch: 2699/22161, Loss=1.6363, lr=0.0000183 Time cost=82.1 Thoughput=14.62 samples/s
INFO:gluonnlp:06:23:14 Epoch: 0, Batch: 2999/22161, Loss=1.5118, lr=0.0000203 Time cost=83.5 Thoughput=14.38 samples/s
INFO:gluonnlp:06:24:36 Epoch: 0, Batch: 3299/22161, Loss=1.4604, lr=0.0000224 Time cost=82.1 Thoughput=14.61 samples/s
INFO:gluonnlp:06:26:00 Epoch: 0, Batch: 3599/22161, Loss=1.4829, lr=0.0000244 Time cost=84.0 Thoughput=14.28 samples/s
INFO:gluonnlp:06:27:20 Epoch: 0, Batch: 3899/22161, Loss=1.3743, lr=0.0000264 Time cost=79.1 Thoughput=15.16 samples/s
INFO:gluonnlp:06:28:45 Epoch: 0, Batch: 4199/22161, Loss=1.3824, lr=0.0000285 Time cost=85.8 Thoughput=13.99 samples/s
INFO:gluonnlp:06:30:09 Epoch: 0, Batch: 4499/22161, Loss=1.3453, lr=0.0000299 Time cost=83.3 Thoughput=14.41 samples/s
INFO:gluonnlp:06:31:33 Epoch: 0, Batch: 4799/22161, Loss=1.3310, lr=0.0000297 Time cost=84.5 Thoughput=14.21 samples/s
INFO:gluonnlp:06:32:56 Epoch: 0, Batch: 5099/22161, Loss=1.2498, lr=0.0000295 Time cost=83.1 Thoughput=14.44 samples/s
INFO:gluonnlp:06:34:20 Epoch: 0, Batch: 5399/22161, Loss=1.1734, lr=0.0000293 Time cost=84.0 Thoughput=14.28 samples/s
INFO:gluonnlp:06:35:42 Epoch: 0, Batch: 5699/22161, Loss=1.1812, lr=0.0000290 Time cost=81.8 Thoughput=14.67 samples/s
INFO:gluonnlp:06:37:06 Epoch: 0, Batch: 5999/22161, Loss=1.1755, lr=0.0000288 Time cost=83.7 Thoughput=14.34 samples/s
INFO:gluonnlp:06:38:27 Epoch: 0, Batch: 6299/22161, Loss=1.2274, lr=0.0000286 Time cost=81.1 Thoughput=14.80 samples/s
INFO:gluonnlp:06:39:51 Epoch: 0, Batch: 6599/22161, Loss=1.2411, lr=0.0000284 Time cost=84.3 Thoughput=14.24 samples/s
INFO:gluonnlp:06:41:14 Epoch: 0, Batch: 6899/22161, Loss=1.1604, lr=0.0000281 Time cost=82.8 Thoughput=14.49 samples/s
INFO:gluonnlp:06:42:36 Epoch: 0, Batch: 7199/22161, Loss=1.2045, lr=0.0000279 Time cost=82.1 Thoughput=14.61 samples/s
INFO:gluonnlp:06:43:58 Epoch: 0, Batch: 7499/22161, Loss=1.1101, lr=0.0000277 Time cost=82.2 Thoughput=14.60 samples/s
INFO:gluonnlp:06:45:22 Epoch: 0, Batch: 7799/22161, Loss=1.1529, lr=0.0000275 Time cost=83.6 Thoughput=14.35 samples/s
INFO:gluonnlp:06:46:48 Epoch: 0, Batch: 8099/22161, Loss=1.0857, lr=0.0000272 Time cost=85.8 Thoughput=13.99 samples/s
INFO:gluonnlp:06:48:06 Epoch: 0, Batch: 8399/22161, Loss=1.1237, lr=0.0000270 Time cost=78.7 Thoughput=15.25 samples/s
INFO:gluonnlp:06:49:29 Epoch: 0, Batch: 8699/22161, Loss=1.0934, lr=0.0000268 Time cost=83.3 Thoughput=14.41 samples/s
INFO:gluonnlp:06:50:55 Epoch: 0, Batch: 8999/22161, Loss=1.1091, lr=0.0000266 Time cost=85.7 Thoughput=14.01 samples/s
INFO:gluonnlp:06:52:17 Epoch: 0, Batch: 9299/22161, Loss=1.1186, lr=0.0000263 Time cost=82.2 Thoughput=14.60 samples/s
INFO:gluonnlp:06:53:40 Epoch: 0, Batch: 9599/22161, Loss=1.0320, lr=0.0000261 Time cost=83.1 Thoughput=14.44 samples/s
INFO:gluonnlp:06:55:01 Epoch: 0, Batch: 9899/22161, Loss=1.0716, lr=0.0000259 Time cost=80.7 Thoughput=14.87 samples/s
INFO:gluonnlp:06:56:24 Epoch: 0, Batch: 10199/22161, Loss=1.0220, lr=0.0000257 Time cost=83.1 Thoughput=14.45 samples/s
INFO:gluonnlp:06:57:47 Epoch: 0, Batch: 10499/22161, Loss=1.0985, lr=0.0000254 Time cost=82.3 Thoughput=14.59 samples/s
INFO:gluonnlp:06:59:05 Epoch: 0, Batch: 10799/22161, Loss=1.0318, lr=0.0000252 Time cost=78.9 Thoughput=15.21 samples/s
INFO:gluonnlp:07:00:30 Epoch: 0, Batch: 11099/22161, Loss=1.1125, lr=0.0000250 Time cost=84.2 Thoughput=14.25 samples/s
INFO:gluonnlp:07:01:52 Epoch: 0, Batch: 11399/22161, Loss=1.0080, lr=0.0000248 Time cost=82.4 Thoughput=14.56 samples/s
INFO:gluonnlp:07:03:14 Epoch: 0, Batch: 11699/22161, Loss=1.0713, lr=0.0000245 Time cost=82.0 Thoughput=14.63 samples/s
INFO:gluonnlp:07:04:35 Epoch: 0, Batch: 11999/22161, Loss=1.0849, lr=0.0000243 Time cost=80.9 Thoughput=14.84 samples/s
INFO:gluonnlp:07:05:57 Epoch: 0, Batch: 12299/22161, Loss=0.9961, lr=0.0000241 Time cost=82.2 Thoughput=14.60 samples/s
INFO:gluonnlp:07:07:18 Epoch: 0, Batch: 12599/22161, Loss=1.0490, lr=0.0000239 Time cost=80.5 Thoughput=14.90 samples/s
INFO:gluonnlp:07:08:40 Epoch: 0, Batch: 12899/22161, Loss=1.0644, lr=0.0000236 Time cost=82.8 Thoughput=14.50 samples/s
INFO:gluonnlp:07:10:01 Epoch: 0, Batch: 13199/22161, Loss=1.0477, lr=0.0000234 Time cost=80.3 Thoughput=14.94 samples/s
INFO:gluonnlp:07:11:24 Epoch: 0, Batch: 13499/22161, Loss=0.9151, lr=0.0000232 Time cost=83.5 Thoughput=14.38 samples/s
INFO:gluonnlp:07:12:50 Epoch: 0, Batch: 13799/22161, Loss=0.9913, lr=0.0000230 Time cost=86.1 Thoughput=13.94 samples/s
INFO:gluonnlp:07:14:15 Epoch: 0, Batch: 14099/22161, Loss=0.9375, lr=0.0000227 Time cost=84.4 Thoughput=14.22 samples/s
INFO:gluonnlp:07:15:39 Epoch: 0, Batch: 14399/22161, Loss=0.9752, lr=0.0000225 Time cost=84.1 Thoughput=14.27 samples/s
INFO:gluonnlp:07:17:06 Epoch: 0, Batch: 14699/22161, Loss=1.0199, lr=0.0000223 Time cost=87.7 Thoughput=13.68 samples/s
INFO:gluonnlp:07:18:30 Epoch: 0, Batch: 14999/22161, Loss=0.9873, lr=0.0000220 Time cost=83.6 Thoughput=14.36 samples/s
INFO:gluonnlp:07:19:55 Epoch: 0, Batch: 15299/22161, Loss=0.9806, lr=0.0000218 Time cost=85.3 Thoughput=14.08 samples/s
INFO:gluonnlp:07:21:16 Epoch: 0, Batch: 15599/22161, Loss=1.0056, lr=0.0000216 Time cost=80.8 Thoughput=14.85 samples/s
INFO:gluonnlp:07:22:36 Epoch: 0, Batch: 15899/22161, Loss=0.9819, lr=0.0000214 Time cost=80.0 Thoughput=15.00 samples/s
INFO:gluonnlp:07:24:00 Epoch: 0, Batch: 16199/22161, Loss=0.9413, lr=0.0000211 Time cost=83.9 Thoughput=14.31 samples/s
INFO:gluonnlp:07:25:21 Epoch: 0, Batch: 16499/22161, Loss=0.9609, lr=0.0000209 Time cost=80.7 Thoughput=14.86 samples/s
INFO:gluonnlp:07:26:41 Epoch: 0, Batch: 16799/22161, Loss=0.9961, lr=0.0000207 Time cost=80.6 Thoughput=14.89 samples/s
INFO:gluonnlp:07:28:03 Epoch: 0, Batch: 17099/22161, Loss=0.8957, lr=0.0000205 Time cost=82.1 Thoughput=14.61 samples/s
INFO:gluonnlp:07:29:29 Epoch: 0, Batch: 17399/22161, Loss=0.8625, lr=0.0000202 Time cost=85.9 Thoughput=13.98 samples/s
INFO:gluonnlp:07:30:52 Epoch: 0, Batch: 17699/22161, Loss=0.9809, lr=0.0000200 Time cost=82.6 Thoughput=14.52 samples/s
INFO:gluonnlp:07:32:15 Epoch: 0, Batch: 17999/22161, Loss=0.8771, lr=0.0000198 Time cost=83.3 Thoughput=14.41 samples/s
INFO:gluonnlp:07:33:38 Epoch: 0, Batch: 18299/22161, Loss=0.8975, lr=0.0000196 Time cost=82.9 Thoughput=14.47 samples/s
INFO:gluonnlp:07:35:01 Epoch: 0, Batch: 18599/22161, Loss=0.9418, lr=0.0000193 Time cost=83.0 Thoughput=14.47 samples/s
INFO:gluonnlp:07:36:23 Epoch: 0, Batch: 18899/22161, Loss=0.8763, lr=0.0000191 Time cost=82.2 Thoughput=14.60 samples/s
INFO:gluonnlp:07:37:48 Epoch: 0, Batch: 19199/22161, Loss=0.9534, lr=0.0000189 Time cost=84.6 Thoughput=14.18 samples/s
INFO:gluonnlp:07:39:09 Epoch: 0, Batch: 19499/22161, Loss=0.8999, lr=0.0000187 Time cost=81.5 Thoughput=14.72 samples/s
INFO:gluonnlp:07:40:31 Epoch: 0, Batch: 19799/22161, Loss=0.9270, lr=0.0000184 Time cost=81.8 Thoughput=14.66 samples/s
INFO:gluonnlp:07:41:56 Epoch: 0, Batch: 20099/22161, Loss=0.9886, lr=0.0000182 Time cost=84.2 Thoughput=14.24 samples/s
INFO:gluonnlp:07:43:18 Epoch: 0, Batch: 20399/22161, Loss=1.0070, lr=0.0000180 Time cost=82.5 Thoughput=14.54 samples/s
INFO:gluonnlp:07:44:41 Epoch: 0, Batch: 20699/22161, Loss=0.9042, lr=0.0000178 Time cost=83.2 Thoughput=14.43 samples/s
INFO:gluonnlp:07:46:05 Epoch: 0, Batch: 20999/22161, Loss=0.9280, lr=0.0000175 Time cost=83.9 Thoughput=14.31 samples/s
INFO:gluonnlp:07:47:27 Epoch: 0, Batch: 21299/22161, Loss=0.9085, lr=0.0000173 Time cost=81.5 Thoughput=14.72 samples/s
INFO:gluonnlp:07:48:51 Epoch: 0, Batch: 21599/22161, Loss=0.7971, lr=0.0000171 Time cost=84.8 Thoughput=14.14 samples/s
INFO:gluonnlp:07:50:15 Epoch: 0, Batch: 21899/22161, Loss=0.9080, lr=0.0000169 Time cost=83.9 Thoughput=14.30 samples/s
INFO:gluonnlp:07:51:24 Time cost=6124.94 s, Thoughput=14.47 samples/s
INFO:gluonnlp:07:52:46 Epoch: 1, Batch: 299/22161, Loss=0.6843, lr=0.0000164 Time cost=81.6 Thoughput=27.47 samples/s
INFO:gluonnlp:07:54:09 Epoch: 1, Batch: 599/22161, Loss=0.6391, lr=0.0000162 Time cost=83.5 Thoughput=14.37 samples/s
INFO:gluonnlp:07:55:30 Epoch: 1, Batch: 899/22161, Loss=0.6156, lr=0.0000160 Time cost=81.0 Thoughput=14.81 samples/s
INFO:gluonnlp:07:56:56 Epoch: 1, Batch: 1199/22161, Loss=0.6392, lr=0.0000158 Time cost=85.8 Thoughput=13.99 samples/s
INFO:gluonnlp:07:58:18 Epoch: 1, Batch: 1499/22161, Loss=0.6291, lr=0.0000155 Time cost=82.4 Thoughput=14.56 samples/s
INFO:gluonnlp:07:59:42 Epoch: 1, Batch: 1799/22161, Loss=0.7006, lr=0.0000153 Time cost=84.0 Thoughput=14.28 samples/s
INFO:gluonnlp:08:01:08 Epoch: 1, Batch: 2099/22161, Loss=0.6279, lr=0.0000151 Time cost=85.1 Thoughput=14.10 samples/s
INFO:gluonnlp:08:02:31 Epoch: 1, Batch: 2399/22161, Loss=0.6512, lr=0.0000149 Time cost=83.0 Thoughput=14.45 samples/s
INFO:gluonnlp:08:03:56 Epoch: 1, Batch: 2699/22161, Loss=0.6549, lr=0.0000146 Time cost=85.5 Thoughput=14.03 samples/s
INFO:gluonnlp:08:05:18 Epoch: 1, Batch: 2999/22161, Loss=0.6766, lr=0.0000144 Time cost=82.0 Thoughput=14.63 samples/s
INFO:gluonnlp:08:06:41 Epoch: 1, Batch: 3299/22161, Loss=0.6206, lr=0.0000142 Time cost=82.5 Thoughput=14.55 samples/s
INFO:gluonnlp:08:08:05 Epoch: 1, Batch: 3599/22161, Loss=0.6309, lr=0.0000140 Time cost=84.2 Thoughput=14.25 samples/s
INFO:gluonnlp:08:09:27 Epoch: 1, Batch: 3899/22161, Loss=0.6020, lr=0.0000137 Time cost=82.1 Thoughput=14.62 samples/s
INFO:gluonnlp:08:10:48 Epoch: 1, Batch: 4199/22161, Loss=0.6479, lr=0.0000135 Time cost=81.3 Thoughput=14.75 samples/s
INFO:gluonnlp:08:12:14 Epoch: 1, Batch: 4499/22161, Loss=0.7397, lr=0.0000133 Time cost=85.7 Thoughput=14.01 samples/s
INFO:gluonnlp:08:13:34 Epoch: 1, Batch: 4799/22161, Loss=0.6255, lr=0.0000131 Time cost=79.7 Thoughput=15.05 samples/s
INFO:gluonnlp:08:14:54 Epoch: 1, Batch: 5099/22161, Loss=0.6023, lr=0.0000128 Time cost=80.6 Thoughput=14.90 samples/s
INFO:gluonnlp:08:16:17 Epoch: 1, Batch: 5399/22161, Loss=0.6514, lr=0.0000126 Time cost=82.8 Thoughput=14.49 samples/s
INFO:gluonnlp:08:17:40 Epoch: 1, Batch: 5699/22161, Loss=0.6869, lr=0.0000124 Time cost=82.5 Thoughput=14.54 samples/s
INFO:gluonnlp:08:19:03 Epoch: 1, Batch: 5999/22161, Loss=0.6292, lr=0.0000121 Time cost=83.3 Thoughput=14.40 samples/s
INFO:gluonnlp:08:20:25 Epoch: 1, Batch: 6299/22161, Loss=0.6005, lr=0.0000119 Time cost=82.4 Thoughput=14.56 samples/s
INFO:gluonnlp:08:21:50 Epoch: 1, Batch: 6599/22161, Loss=0.6261, lr=0.0000117 Time cost=84.9 Thoughput=14.13 samples/s
INFO:gluonnlp:08:23:14 Epoch: 1, Batch: 6899/22161, Loss=0.6152, lr=0.0000115 Time cost=83.8 Thoughput=14.32 samples/s
INFO:gluonnlp:08:24:35 Epoch: 1, Batch: 7199/22161, Loss=0.6314, lr=0.0000112 Time cost=80.7 Thoughput=14.86 samples/s
INFO:gluonnlp:08:25:58 Epoch: 1, Batch: 7499/22161, Loss=0.6823, lr=0.0000110 Time cost=83.3 Thoughput=14.40 samples/s
INFO:gluonnlp:08:27:22 Epoch: 1, Batch: 7799/22161, Loss=0.6816, lr=0.0000108 Time cost=83.8 Thoughput=14.32 samples/s
INFO:gluonnlp:08:28:44 Epoch: 1, Batch: 8099/22161, Loss=0.6695, lr=0.0000106 Time cost=82.4 Thoughput=14.57 samples/s
INFO:gluonnlp:08:30:08 Epoch: 1, Batch: 8399/22161, Loss=0.6349, lr=0.0000103 Time cost=83.5 Thoughput=14.37 samples/s
INFO:gluonnlp:08:31:30 Epoch: 1, Batch: 8699/22161, Loss=0.6403, lr=0.0000101 Time cost=82.7 Thoughput=14.51 samples/s
INFO:gluonnlp:08:32:55 Epoch: 1, Batch: 8999/22161, Loss=0.6395, lr=0.0000099 Time cost=85.0 Thoughput=14.12 samples/s
INFO:gluonnlp:08:34:20 Epoch: 1, Batch: 9299/22161, Loss=0.6655, lr=0.0000097 Time cost=85.0 Thoughput=14.12 samples/s
INFO:gluonnlp:08:35:41 Epoch: 1, Batch: 9599/22161, Loss=0.6224, lr=0.0000094 Time cost=80.9 Thoughput=14.83 samples/s
INFO:gluonnlp:08:37:03 Epoch: 1, Batch: 9899/22161, Loss=0.6160, lr=0.0000092 Time cost=81.9 Thoughput=14.65 samples/s
INFO:gluonnlp:08:38:26 Epoch: 1, Batch: 10199/22161, Loss=0.6879, lr=0.0000090 Time cost=82.7 Thoughput=14.52 samples/s
INFO:gluonnlp:08:39:48 Epoch: 1, Batch: 10499/22161, Loss=0.6595, lr=0.0000088 Time cost=81.8 Thoughput=14.67 samples/s
INFO:gluonnlp:08:41:11 Epoch: 1, Batch: 10799/22161, Loss=0.6826, lr=0.0000085 Time cost=83.0 Thoughput=14.45 samples/s
INFO:gluonnlp:08:42:30 Epoch: 1, Batch: 11099/22161, Loss=0.6699, lr=0.0000083 Time cost=79.4 Thoughput=15.12 samples/s
INFO:gluonnlp:08:43:54 Epoch: 1, Batch: 11399/22161, Loss=0.6629, lr=0.0000081 Time cost=84.4 Thoughput=14.22 samples/s
INFO:gluonnlp:08:45:17 Epoch: 1, Batch: 11699/22161, Loss=0.5918, lr=0.0000079 Time cost=82.2 Thoughput=14.60 samples/s
INFO:gluonnlp:08:46:38 Epoch: 1, Batch: 11999/22161, Loss=0.6440, lr=0.0000076 Time cost=81.8 Thoughput=14.67 samples/s
INFO:gluonnlp:08:48:02 Epoch: 1, Batch: 12299/22161, Loss=0.6274, lr=0.0000074 Time cost=83.6 Thoughput=14.35 samples/s
INFO:gluonnlp:08:49:24 Epoch: 1, Batch: 12599/22161, Loss=0.6499, lr=0.0000072 Time cost=82.4 Thoughput=14.57 samples/s
INFO:gluonnlp:08:50:47 Epoch: 1, Batch: 12899/22161, Loss=0.6308, lr=0.0000070 Time cost=82.2 Thoughput=14.60 samples/s
INFO:gluonnlp:08:52:11 Epoch: 1, Batch: 13199/22161, Loss=0.6407, lr=0.0000067 Time cost=84.0 Thoughput=14.29 samples/s
INFO:gluonnlp:08:53:34 Epoch: 1, Batch: 13499/22161, Loss=0.6211, lr=0.0000065 Time cost=82.9 Thoughput=14.48 samples/s
INFO:gluonnlp:08:54:56 Epoch: 1, Batch: 13799/22161, Loss=0.5870, lr=0.0000063 Time cost=82.9 Thoughput=14.48 samples/s
INFO:gluonnlp:08:56:18 Epoch: 1, Batch: 14099/22161, Loss=0.6334, lr=0.0000061 Time cost=82.0 Thoughput=14.64 samples/s
INFO:gluonnlp:08:57:39 Epoch: 1, Batch: 14399/22161, Loss=0.6223, lr=0.0000058 Time cost=80.9 Thoughput=14.84 samples/s
INFO:gluonnlp:08:59:00 Epoch: 1, Batch: 14699/22161, Loss=0.6175, lr=0.0000056 Time cost=80.6 Thoughput=14.89 samples/s
INFO:gluonnlp:09:00:24 Epoch: 1, Batch: 14999/22161, Loss=0.6223, lr=0.0000054 Time cost=84.5 Thoughput=14.20 samples/s
INFO:gluonnlp:09:01:48 Epoch: 1, Batch: 15299/22161, Loss=0.6447, lr=0.0000052 Time cost=83.8 Thoughput=14.32 samples/s
INFO:gluonnlp:09:03:14 Epoch: 1, Batch: 15599/22161, Loss=0.5361, lr=0.0000049 Time cost=85.7 Thoughput=14.00 samples/s
INFO:gluonnlp:09:04:34 Epoch: 1, Batch: 15899/22161, Loss=0.6136, lr=0.0000047 Time cost=80.4 Thoughput=14.92 samples/s
INFO:gluonnlp:09:05:57 Epoch: 1, Batch: 16199/22161, Loss=0.5523, lr=0.0000045 Time cost=82.4 Thoughput=14.57 samples/s
INFO:gluonnlp:09:07:18 Epoch: 1, Batch: 16499/22161, Loss=0.6702, lr=0.0000043 Time cost=81.8 Thoughput=14.67 samples/s
INFO:gluonnlp:09:08:43 Epoch: 1, Batch: 16799/22161, Loss=0.5886, lr=0.0000040 Time cost=84.9 Thoughput=14.14 samples/s
INFO:gluonnlp:09:10:08 Epoch: 1, Batch: 17099/22161, Loss=0.6240, lr=0.0000038 Time cost=84.7 Thoughput=14.16 samples/s
INFO:gluonnlp:09:11:32 Epoch: 1, Batch: 17399/22161, Loss=0.5796, lr=0.0000036 Time cost=83.8 Thoughput=14.32 samples/s
INFO:gluonnlp:09:12:55 Epoch: 1, Batch: 17699/22161, Loss=0.5557, lr=0.0000033 Time cost=83.6 Thoughput=14.35 samples/s
INFO:gluonnlp:09:14:19 Epoch: 1, Batch: 17999/22161, Loss=0.5857, lr=0.0000031 Time cost=83.5 Thoughput=14.37 samples/s
INFO:gluonnlp:09:15:43 Epoch: 1, Batch: 18299/22161, Loss=0.6348, lr=0.0000029 Time cost=83.9 Thoughput=14.31 samples/s
INFO:gluonnlp:09:17:03 Epoch: 1, Batch: 18599/22161, Loss=0.6214, lr=0.0000027 Time cost=79.8 Thoughput=15.05 samples/s
INFO:gluonnlp:09:18:23 Epoch: 1, Batch: 18899/22161, Loss=0.6367, lr=0.0000024 Time cost=80.5 Thoughput=14.91 samples/s
INFO:gluonnlp:09:19:45 Epoch: 1, Batch: 19199/22161, Loss=0.6122, lr=0.0000022 Time cost=82.2 Thoughput=14.60 samples/s
INFO:gluonnlp:09:21:07 Epoch: 1, Batch: 19499/22161, Loss=0.6111, lr=0.0000020 Time cost=81.7 Thoughput=14.69 samples/s
INFO:gluonnlp:09:22:29 Epoch: 1, Batch: 19799/22161, Loss=0.6016, lr=0.0000018 Time cost=81.6 Thoughput=14.71 samples/s
INFO:gluonnlp:09:23:51 Epoch: 1, Batch: 20099/22161, Loss=0.6507, lr=0.0000015 Time cost=82.9 Thoughput=14.47 samples/s
INFO:gluonnlp:09:25:14 Epoch: 1, Batch: 20399/22161, Loss=0.6095, lr=0.0000013 Time cost=82.8 Thoughput=14.49 samples/s
INFO:gluonnlp:09:26:36 Epoch: 1, Batch: 20699/22161, Loss=0.6100, lr=0.0000011 Time cost=82.2 Thoughput=14.60 samples/s
INFO:gluonnlp:09:27:59 Epoch: 1, Batch: 20999/22161, Loss=0.6042, lr=0.0000009 Time cost=82.3 Thoughput=14.58 samples/s
INFO:gluonnlp:09:29:21 Epoch: 1, Batch: 21299/22161, Loss=0.6265, lr=0.0000006 Time cost=82.7 Thoughput=14.50 samples/s
INFO:gluonnlp:09:30:45 Epoch: 1, Batch: 21599/22161, Loss=0.6006, lr=0.0000004 Time cost=83.1 Thoughput=14.44 samples/s
INFO:gluonnlp:09:32:07 Epoch: 1, Batch: 21899/22161, Loss=0.5957, lr=0.0000002 Time cost=82.6 Thoughput=14.52 samples/s
INFO:gluonnlp:09:33:16 Finish training step: 7386
INFO:gluonnlp:09:33:16 Time cost=12236.99 s, Thoughput=14.48 samples/s
INFO:gluonnlp:09:33:26 Loading dev data...
INFO:gluonnlp:09:33:27 Number of records in dev data:10570
INFO:gluonnlp:09:33:38 The number of examples after preprocessing:10833
INFO:gluonnlp:09:33:39 start prediction
INFO:gluonnlp:09:36:39 Time cost=180.52 s, Thoughput=60.01 samples/s
INFO:gluonnlp:09:36:39 Get prediction results...
INFO:gluonnlp:09:37:23 {'exact_match': 84.22894985808892, 'f1': 90.96706973833493}