marsggbo/bigbench4switch64_patternand_pattern_predictor_gen

--- dataset_info: features: - name: prompt_text dtype: string - name: prompt_ids sequence: int64 - name: decode_ids sequence: int64 - name: prompt_pattern sequence: sequence: int64 - name: decode_pattern sequence: sequence: int64 - name: predictor_pattern sequence: sequence: sequence: int64 splits: - name: train num_bytes: 219905957 num_examples: 10936 download_size: 23650636 dataset_size: 219905957 configs: - config_name: default data_files: - split: train path: data/train-* ---The dataset includes multiple features such as prompt_text (string), prompt_ids (sequence of int64), decode_ids (sequence of int64), prompt_pattern (sequence of sequence of int64), decode_pattern (sequence of sequence of int64), and predictor_pattern (sequence of sequence of sequence of int64). The dataset is split into a training set (train) with 10936 examples, totaling 219905957 bytes. The download size of the dataset is 23650636 bytes.

原始信息汇总

数据集概述

数据集特征

• prompt_text：字符串类型
• prompt_ids：序列类型，整数64位
• decode_ids：序列类型，整数64位
• prompt_pattern：序列类型，内部序列类型为整数64位
• decode_pattern：序列类型，内部序列类型为整数64位
• predictor_pattern：序列类型，内部序列类型为整数64位

数据集划分

• train：
  • 数据量：219,905,957字节
  • 示例数量：10,936

数据集大小

• 下载大小：23,650,636字节
• 数据集大小：219,905,957字节

配置信息

• config_name: default
• data_files:
  • split: train
  • path: data/train-*