search

HRNet / HRNet-Object-Detection

Object detection with multi-level representations generated from deep high-resolution representation learning (HRNetV2h). This is an official implementation for our TPAMI paper "Deep High-Resolution Representation Learning for Visual Recognition". https://arxiv.org/abs/1908.07919
Apache License 2.0
644 stars 97 forks source link

the first iteration, acc: 94.7188, is it normal? #13

Open passion3394 opened 5 years ago

passion3394 commented 5 years ago

2019-07-08 19:39:57,836 - INFO - Epoch [1][50/39274] lr: 0.00040, eta: 2 days, 23:25:49, time: 1.091, data_time: 0.013, memory: 4931, loss_rpn_cls: 0.6880, loss_rpn_reg: 0.0123, loss_cls: 3.2936, acc: 94.7188, loss_reg: 0.0006, loss: 3.9945 2019-07-08 19:40:50,242 - INFO - Epoch [1][100/39274] lr: 0.00047, eta: 2 days, 22:01:12, time: 1.049, data_time: 0.005, memory: 4931, loss_rpn_cls: 0.6152, loss_rpn_reg: 0.0125, loss_cls: 1.0380, acc: 99.7266, loss_reg: 0.0000, loss: 1.6657 2019-07-08 19:41:42,772 - INFO - Epoch [1][150/39274] lr: 0.00053, eta: 2 days, 21:34:41, time: 1.051, data_time: 0.004, memory: 4931, loss_rpn_cls: 0.2101, loss_rpn_reg: 0.0120, loss_cls: 0.1034, acc: 99.7227, loss_reg: 0.0000, loss: 0.3256 2019-07-08 19:42:35,451 - INFO - Epoch [1][200/39274] lr: 0.00060, eta: 2 days, 21:23:54, time: 1.054, data_time: 0.004, memory: 4931, loss_rpn_cls: 0.1267, loss_rpn_reg: 0.0133, loss_cls: 0.0715, acc: 99.7305, loss_reg: 0.0001, loss: 0.2116 2019-07-08 19:43:28,060 - INFO - Epoch [1][250/39274] lr: 0.00067, eta: 2 days, 21:16:00, time: 1.052, data_time: 0.004, memory: 4931, loss_rpn_cls: 0.0997, loss_rpn_reg: 0.0096, loss_cls: 0.0475, acc: 99.7344, loss_reg: 0.0000, loss: 0.1568 2019-07-08 19:44:20,587 - INFO - Epoch [1][300/39274] lr: 0.00073, eta: 2 days, 21:09:22, time: 1.051, data_time: 0.004, memory: 4931, loss_rpn_cls: 0.1093, loss_rpn_reg: 0.0097, loss_cls: 0.0364, acc: 99.7188, loss_reg: 0.0000, loss: 0.1554 2019-07-08 19:45:13,128 - INFO - Epoch [1][350/39274] lr: 0.00080, eta: 2 days, 21:04:31, time: 1.051, data_time: 0.004, memory: 4931, loss_rpn_cls: 0.1045, loss_rpn_reg: 0.0087, loss_cls: 0.0325, acc: 99.7031, loss_reg: 0.0000, loss: 0.1456 2019-07-08 19:46:05,796 - INFO - Epoch [1][400/39274] lr: 0.00087, eta: 2 days, 21:01:55, time: 1.053, data_time: 0.004, memory: 4931, loss_rpn_cls: 0.1136, loss_rpn_reg: 0.0069, loss_cls: 0.0339, acc: 99.6758, loss_reg: 0.0004, loss: 0.1547 2019-07-08 19:46:58,322 - INFO - Epoch [1][450/39274] lr: 0.00093, eta: 2 days, 20:58:28, time: 1.051, data_time: 0.004, memory: 4931, loss_rpn_cls: 0.1078, loss_rpn_reg: 0.0074, loss_cls: 0.0314, acc: 99.7031, loss_reg: 0.0000, loss: 0.1465 2019-07-08 19:47:50,988 - INFO - Epoch [1][500/39274] lr: 0.00100, eta: 2 days, 20:56:38, time: 1.053, data_time: 0.004, memory: 4931, loss_rpn_cls: 0.0997, loss_rpn_reg: 0.0073, loss_cls: 0.0296, acc: 99.7188, loss_reg: 0.0000, loss: 0.1366 2019-07-08 19:48:43,311 - INFO - Epoch [1][550/39274] lr: 0.00100, eta: 2 days, 20:52:31, time: 1.046, data_time: 0.004, memory: 4931, loss_rpn_cls: 0.1149, loss_rpn_reg: 0.0110, loss_cls: 0.0291, acc: 99.7148, loss_reg: 0.0000, loss: 0.1551 2019-07-08 19:49:35,827 - INFO - Epoch [1][600/39274] lr: 0.00100, eta: 2 days, 20:50:12, time: 1.050, data_time: 0.004, memory: 4931, loss_rpn_cls: 0.1092, loss_rpn_reg: 0.0094, loss_cls: 0.0275, acc: 99.7305, loss_reg: 0.0000, loss: 0.1461 2019-07-08 19:50:28,301 - INFO - Epoch [1][650/39274] lr: 0.00100, eta: 2 days, 20:47:52, time: 1.049, data_time: 0.004, memory: 4931, loss_rpn_cls: 0.0997, loss_rpn_reg: 0.0084, loss_cls: 0.0271, acc: 99.7188, loss_reg: 0.0000, loss: 0.1353 2019-07-08 19:51:20,748 - INFO - Epoch [1][700/39274] lr: 0.00100, eta: 2 days, 20:45:35, time: 1.049, data_time: 0.004, memory: 4931, loss_rpn_cls: 0.1172, loss_rpn_reg: 0.0104, loss_cls: 0.0309, acc: 99.6719, loss_reg: 0.0000, loss: 0.1585 2019-07-08 19:52:13,283 - INFO - Epoch [1][750/39274] lr: 0.00100, eta: 2 days, 20:43:57, time: 1.051, data_time: 0.004, memory: 4931, loss_rpn_cls: 0.1072, loss_rpn_reg: 0.0108, loss_cls: 0.0262, acc: 99.7227, loss_reg: 0.0005, loss: 0.1447 2019-07-08 19:53:05,903 - INFO - Epoch [1][800/39274] lr: 0.00100, eta: 2 days, 20:42:49, time: 1.052, data_time: 0.008, memory: 4931, loss_rpn_cls: 0.0828, loss_rpn_reg: 0.0061, loss_cls: 0.0236, acc: 99.7383, loss_reg: 0.0000, loss: 0.1125 2019-07-08 19:53:58,468 - INFO - Epoch [1][850/39274] lr: 0.00100, eta: 2 days, 20:41:28, time: 1.051, data_time: 0.004, memory: 4931, loss_rpn_cls: 0.0966, loss_rpn_reg: 0.0070, loss_cls: 0.0248, acc: 99.7266, loss_reg: 0.0000, loss: 0.1284 2019-07-08 19:54:51,062 - INFO - Epoch [1][900/39274] lr: 0.00100, eta: 2 days, 20:40:18, time: 1.052, data_time: 0.005, memory: 4931, loss_rpn_cls: 0.0997, loss_rpn_reg: 0.0058, loss_cls: 0.0253, acc: 99.7227, loss_reg: 0.0001, loss: 0.1309 2019-07-08 19:55:43,428 - INFO - Epoch [1][950/39274] lr: 0.00100, eta: 2 days, 20:38:13, time: 1.047, data_time: 0.004, memory: 4931, loss_rpn_cls: 0.0916, loss_rpn_reg: 0.0074, loss_cls: 0.0267, acc: 99.6992, loss_reg: 0.0000, loss: 0.1257

wondervictor commented 5 years ago

normal you can refer to the validation results of the first epoch.

passion3394 commented 5 years ago

@wondervictor thanks for your reply. I have two more questions: (1)what's the acc mean? (2)all the loss values are very small from the beginning of the training, should I tune some place to make the losses much higher?