昨天的模型跑了8个小时都没有结束，训练它时用的参数就是文档里面的参数，模型训练的结果如下：
loading annotations into memory... Done (t=0.09s) creating index... index created! [04/06 07:55:34] ppdet.metrics.coco_utils INFO: Start evaluate... Loading and preparing results... DONE (t=15.93s) creating index... index created! Running per image evaluation... Evaluate annotation type *bbox* DONE (t=75.35s). Accumulating evaluation results... DONE (t=17.80s). Average Precision (AP) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.725 Average Precision (AP) @[ IoU=0.50 | area= all | maxDets=100 ] = 0.936 Average Precision (AP) @[ IoU=0.75 | area= all | maxDets=100 ] = 0.824 Average Precision (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.484 Average Precision (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.657 Average Precision (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.847 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 1 ] = 0.348 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 10 ] = 0.751 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.799 Average Recall (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.611 Average Recall (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.774 Average Recall (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.891 [04/06 07:57:27] ppdet.engine INFO: Total sample number: 5691, averge FPS: 14.699495757040536 [04/06 07:57:27] ppdet.engine INFO: Best test bbox ap is 0.725. [04/06 07:57:32] ppdet.utils.checkpoint INFO: Save checkpoint: output/ppyoloe_plus_crn_l_80e_coco [04/06 07:57:33] ppdet.engine INFO: Epoch: [55] [ 0/711] learning_rate: 0.000053 loss: 1.361498 loss_cls: 0.597509 loss_iou: 0.153269 loss_dfl: 0.733937 loss_l1: 0.224196 eta: 2:14:26 batch_cost: 0.4401 data_cost: 0.0004 ips: 18.1778 images/s [04/06 07:58:30] ppdet.engine INFO: Epoch: [55] [100/711] learning_rate: 0.000053 loss: 1.333801 loss_cls: 0.582086 loss_iou: 0.153208 loss_dfl: 0.721183 loss_l1: 0.208167 eta: 2:13:39 batch_cost: 0.4317 data_cost: 0.0004 ips: 18.5325 images/s [04/06 07:59:29] ppdet.engine INFO: Epoch: [55] [200/711] learning_rate: 0.000052 loss: 1.310558 loss_cls: 0.570390 loss_iou: 0.148457 loss_dfl: 0.723420 loss_l1: 0.218413 eta: 2:12:54 batch_cost: 0.4504 data_cost: 0.0003 ips: 17.7608 images/s [04/06 08:00:26] ppdet.engine INFO: Epoch: [55] [300/711] learning_rate: 0.000052 loss: 1.317477 loss_cls: 0.584557 loss_iou: 0.151085 loss_dfl: 0.709756 loss_l1: 0.203959 eta: 2:12:08 batch_cost: 0.4339 data_cost: 0.0003 ips: 18.4377 images/s [04/06 08:01:21] ppdet.engine INFO: Epoch: [55] [400/711] learning_rate: 0.000052 loss: 1.291385 loss_cls: 0.557138 loss_iou: 0.146287 loss_dfl: 0.705459 loss_l1: 0.199396 eta: 2:11:21 batch_cost: 0.4256 data_cost: 0.0003 ips: 18.7982 images/s [04/06 08:02:21] ppdet.engine INFO: Epoch: [55] [500/711] learning_rate: 0.000051 loss: 1.314564 loss_cls: 0.579978 loss_iou: 0.152582 loss_dfl: 0.726087 loss_l1: 0.214535 eta: 2:10:36 batch_cost: 0.4592 data_cost: 0.0003 ips: 17.4214 images/s [04/06 08:03:13] ppdet.engine INFO: Epoch: [55] [600/711] learning_rate: 0.000051 loss: 1.320678 loss_cls: 0.590191 loss_iou: 0.149505 loss_dfl: 0.711055 loss_l1: 0.199503 eta: 2:09:48 batch_cost: 0.4030 data_cost: 0.0003 ips: 19.8509 images/s [04/06 08:04:11] ppdet.engine INFO: Epoch: [55] [700/711] learning_rate: 0.000051 loss: 1.335601 loss_cls: 0.581962 loss_iou: 0.151998 loss_dfl: 0.725094 loss_l1: 0.224158 eta: 2:09:03 batch_cost: 0.4470 data_cost: 0.0007 ips: 17.8983 images/s [04/06 08:04:21] ppdet.engine INFO: Epoch: [56] [ 0/711] learning_rate: 0.000051 loss: 1.328567 loss_cls: 0.581747 loss_iou: 0.151998 loss_dfl: 0.723408 loss_l1: 0.217804 eta: 2:08:58 batch_cost: 0.4622 data_cost: 0.0116 ips: 17.3090 images/s [04/06 08:05:18] ppdet.engine INFO: Epoch: [56] [100/711] learning_rate: 0.000050 loss: 1.321686 loss_cls: 0.581012 loss_iou: 0.150127 loss_dfl: 0.706825 loss_l1: 0.208181 eta: 2:08:12 batch_cost: 0.4367 data_cost: 0.0003 ips: 18.3178 images/s [04/06 08:06:14] ppdet.engine INFO: Epoch: [56] [200/711] learning_rate: 0.000050 loss: 1.318650 loss_cls: 0.575591 loss_iou: 0.144920 loss_dfl: 0.715922 loss_l1: 0.211983 eta: 2:07:26 batch_cost: 0.4340 data_cost: 0.0003 ips: 18.4337 images/s
解析：
这是一个目标检测模型在测试集上的评估结果，包括了不同IoU阈值、不同物体大小和不同数量的最大检测框数下的平均精度（Average Precision，AP）和平均召回率（Average Recall，AR）。其中IoU指交并比（Intersection over Union），是目标检测中用于衡量预测框和真实框之间重合程度的指标，物体大小分为small、medium和large三种，maxDets表示每张图片最多可以检测到的目标数量。
具体来说，这个模型在测试集上的AP@[IoU=0.5:0.95 | area=all | maxDets=100]为0.725，这个指标可以用来衡量模型的整体性能；AP@[IoU=0.5 | area=all | maxDets=100]、AP@[IoU=0.75 | area=all | maxDets=100]可以分别用来衡量模型在不同IoU阈值下的表现；AP@[IoU=0.5:0.95 | area=small | maxDets=100]、AP@[IoU=0.5:0.95 | area=medium | maxDets=100]、AP@[IoU=0.5:0.95 | area=large | maxDets=100]可以分别用来衡量模型在不同物体大小下的表现；AR@[IoU=0.5:0.95 | area=all | maxDets=1]、AR@[IoU=0.5:0.95 | area=all | maxDets=10]、AR@[IoU=0.5:0.95 | area=all | maxDets=100]可以分别用来衡量模型在不同最大检测框数下的表现。
除此之外，还展示了模型的总样本数、平均每秒处理帧数、最好的测试集bbox ap以及当前训练的进度。其中bbox ap指的是只考虑bounding box的AP，通常用于目标检测中。