Swin-Transformer 图像分类网络实战项目：115种化妆品品牌logo图像分类数据集（迁移学习）

[train hyper-parameters: Namespace(batch_size=16, num_classes=115, lr=0.001, lrf=0.01)] [epoch: 1] train loss:0.2487 test accuracy:0.3585 [epoch: 2] train loss:0.1400 test accuracy:0.5042 [epoch: 3] train loss:0.0901 test accuracy:0.5758 [epoch: 4] train loss:0.0583 test accuracy:0.6116 [epoch: 5] train loss:0.0361 test accuracy:0.6252 [epoch: 6] train loss:0.0224 test accuracy:0.6389 [epoch: 7] train loss:0.0149 test accuracy:0.6487 [epoch: 8] train loss:0.0116 test accuracy:0.6480 [epoch: 9] train loss:0.0096 test accuracy:0.6539 [epoch: 10] train loss:0.0094 test accuracy:0.6545