The training split is used to train models, the validation is used to benchmark different model architectures. The test is ideally held out till the very end ...
確定! 回上一頁