What are the differences between train, val, and test that are often encountered during training in deep learning?

train和val都是为了训练模型参数，test是在参数完全确定后做测试，是衡量你的模型性能。

Train is the training set, and val is the test set during the training process. This is to allow you to see the training results while training and judge the learning status in time. test is the test set used to evaluate the model results after training the model. Only train can be trained, val is not necessary, and the ratio can be set very small. test is not necessary for model training, but some should generally be reserved for testing. The recommended ratio is usually 8:1:1

val都是在训练的时候起作用。而因为val的数据集和train没有交集，所以这部分数据对最终训练出的模型没有贡献。val的主要作用是来验证是否过拟合、以及用来调节训练参数等。