An easy/swift-to-adapt PyTorch-Lighting template. 套壳模板，简单易用，稍改原来Pytorch代码，即可适配Lightning。You can translate your previous Pytorch code much easier using this template, and keep your freedom to edit all the functions as well. Big-project-friendly as well. No need to rewrite your config in hydra.
a897456 opened this issue · comments
loss是在training_step和validation_step产生，但是在main（）中好像没有loss的字眼。请问是怎么back的？
lightning 需要在模型类中定义training_step函数，在training_step函数中进行反向传播，详见文档：https://lightning.ai/docs/pytorch/stable/model/train_model_basic.html