PictoViLT/CG_L4_img_T

该数据集包含以下字段：input_ids（整数序列），attention_mask（字节数组），token_type_ids（字节数组），labels（长整型序列），pixel_values（三维浮点数组），masked_indices（长整型序列），metadata（包含布尔类型和字符串类型的结构）。数据集分为训练集，共有617个示例，总大小为1,095,181,943字节。提供默认配置，数据文件路径指向训练数据。

The dataset includes the following fields: input_ids (integer sequence), attention_mask (byte array), token_type_ids (byte array), labels (long integer sequence), pixel_values (three-dimensional float array), masked_indices (long integer sequence), metadata (containing boolean and string types). The dataset is split into a training set with a total of 617 examples and a size of 1,095,181,943 bytes. A default configuration is provided, pointing to the training data files.