test mAP bigger than 1 during training
Created by: tangxiaochu123230
2020-09-18 13:07:52,170-INFO: Save model to output\faster_rcnn_r101_fpn_1x\30000. 2020-09-18 13:08:00,840-INFO: Test iter 0 2020-09-18 13:08:12,119-INFO: Test iter 100 2020-09-18 13:08:23,475-INFO: Test iter 200 2020-09-18 13:08:28,833-INFO: Test finish iter 248 2020-09-18 13:08:28,833-INFO: Total number of images: 248, inference time: 8.814965413086895 fps. 2020-09-18 13:08:28,833-INFO: Start evaluate... 2020-09-18 13:08:28,896-INFO: Accumulating evaluatation results... 2020-09-18 13:08:28,896-INFO: mAP(0.50, 11point) = 1.33 2020-09-18 13:08:28,896-INFO: Save model to output\faster_rcnn_r101_fpn_1x\best_model. 2020-09-18 13:08:37,363-INFO: Best test box ap: 1.332228025756195, in iter: 30000
2020-09-18 14:06:43,226-INFO: Save model to output\faster_rcnn_r101_fpn_1x\40000. 2020-09-18 14:06:51,864-INFO: Test iter 0 2020-09-18 14:07:03,190-INFO: Test iter 100 2020-09-18 14:07:14,500-INFO: Test iter 200 2020-09-18 14:07:19,858-INFO: Test finish iter 248 2020-09-18 14:07:19,858-INFO: Total number of images: 248, inference time: 8.81987237484895 fps. 2020-09-18 14:07:19,858-INFO: Start evaluate... 2020-09-18 14:07:19,920-INFO: Accumulating evaluatation results... 2020-09-18 14:07:19,936-INFO: mAP(0.50, 11point) = 3.62 2020-09-18 14:07:19,936-INFO: Save model to output\faster_rcnn_r101_fpn_1x\best_model. 2020-09-18 14:07:28,699-INFO: Best test box ap: 3.6219306119180965, in iter: 40000