针对多GPu训练的问题

[xiaoxusanheyi]

1. 执行了你们给出的多GPU训练的命令： 30个epoch 训练 还有10个深度训练 后面直接10次epoch 结果就是直接用测试集测试出来的最终结果吗 我拿到的这个结果是不是就可以和原文进行对比了。
2. 是不是没有用到验证集val呀

[Cc-Hy]

The above commands first use 30 epoch training with BEV feature distillation only to make the student model learn to generate reasonable BEV representations first and then use 10 epoch training with the complete training losses for detection optimization. And the second command will automatically run the eval process on the validation set. And the results given in the model zoo are obtained in the same way, you can compare them fairly.