Training, Validation, and Test Sets: Why We Split Our Data Three Ways
中文摘要
本文探讨了在构建神经网络时，将数据集分为训练集、验证集和测试集的原因。
English Summary
This article explains the reasons for splitting data into training, validation, and test sets when building a neural network.
Original Excerpt
This is Day 16 of building a neural network from scratch. To view previous articles in the series, click here. Continue reading on Medium »