fxwang/easy-math-trees_v2-round-3

该数据集包含用于模型训练和损失计算的输入特征。model_inputs特征包括预排序索引、后排序索引、输入ID、标签、位置ID和序列ID等序列数据。loss_inputs特征包括父索引、正确数量、滚动次数和父滚动索引等序列数据。数据集包含一个训练集分割，共有4461个样本，总大小为237740640字节。

This dataset contains input features for model training and loss calculation. The model_inputs feature includes sequence data such as preorder index, postorder index, input IDs, labels, position IDs, and sequence IDs. The loss_inputs feature includes sequence data such as parent index, number of correct, number of rollouts, and parent rollout index. The dataset includes a training split with 4461 examples and a total size of 237740640 bytes.