Deep Learning Training Data

该数据集包含了来自不同深度学习配置的基准测试结果，这些配置包括全连接层和卷积层，并在不同的硬件平台上进行了测试。数据被均匀分布地划分为训练集（80%）、测试集（10%）和验证集（10%），且特征是采用统一分布进行选择的。该数据集涵盖了全连接层的25,000种参数组合结果以及卷积层的50,000种组合结果。其任务是预测深度学习模型的执行时间。

This dataset comprises benchmark test results obtained from various deep learning configurations, including fully connected layers and convolutional layers, which were evaluated across different hardware platforms. The data is uniformly split into the training set (80%), test set (10%), and validation set (10%), with features selected following a uniform distribution. This dataset includes 25,000 sets of parameter combination results for fully connected layers and 50,000 sets for convolutional layers. The core task of this dataset is to predict the execution time of deep learning models.