README.md

第一次模型训练结果

200 epochs completed in 0.231 hours.
Optimizer stripped from runs\train\exp8\weights\last.pt, 173.1MB
Optimizer stripped from runs\train\exp8\weights\best.pt, 173.1MB

Validating runs\train\exp8\weights\best.pt...
Fusing layers... 
Model summary: 322 layers, 86180143 parameters, 0 gradients, 203.8 GFLOPs
Class Images Instances P R mAP50 mAP50-95: 100%|██████████| 1/1 [00:00<00:00, 3.25it/s]
all 7 95 0.84 0.884 0.877 0.657
luggage 7 55 0.823 0.818 0.832 0.621
person 7 40 0.857 0.95 0.921 0.692
Results saved to runs\train\exp8

测试结果评价

人物检测精度足够，行李箱精度足够，背包，挎包等出现混乱和识别不出的情况

改进措施

1.增加背包挎包手提包等多样性行李的数据集进行训练，2.提高训练轮数。

第二次训练结果

Stopping training early as no improvement observed in last 100 epochs. Best results observed at epoch 185, best model saved as best.pt.
To update EarlyStopping(patience=100) pass a new patience value, i.e. `python train.py --patience 300` or use `--patience 0` to disable EarlyStopping.

286 epochs completed in 0.402 hours.
Optimizer stripped from runs\train\exp\weights\last.pt, 173.1MB
Optimizer stripped from runs\train\exp\weights\best.pt, 173.1MB

Validating runs\train\exp\weights\best.pt...
Fusing layers...
Model summary: 322 layers, 86180143 parameters, 0 gradients, 203.8 GFLOPs
                 Class     Images  Instances          P          R      mAP50   mAP50-95: 100%|██████████| 1/1 [00:00<00:00,  2.70it/s]
                   all          9        129      0.899      0.804      0.864      0.623
               luggage          9         77      0.861      0.766      0.832      0.606
                person          9         52      0.936      0.843      0.896      0.641
Results saved to runs\train\exp

本次训练中，yolo自动终止了训练，因为在过去的100轮中，并没有任何改善，故保存了最佳的第185次