cassanof/minecraft-sample-instructed

--- dataset_info: features: - name: path dtype: string - name: shape sequence: int32 - name: volume dtype: int32 - name: blocks sequence: sequence: sequence: sequence: int32 - name: image dtype: image - name: block_counts sequence: sequence: int32 - name: descriptions sequence: string - name: descriptions_scores sequence: int64 - name: top_description dtype: string splits: - name: train num_bytes: 54208291.0 num_examples: 655 download_size: 14813433 dataset_size: 54208291.0 configs: - config_name: default data_files: - split: train path: data/train-* ---

原始信息汇总

数据集概述

数据集特征

• path: 字符串类型
• shape: 整数序列类型，使用int32
• volume: 整数类型，使用int32
• blocks: 复杂序列类型，最终序列元素为int32
• image: 图像类型
• block_counts: 序列类型，子序列元素为int32
• descriptions: 字符串序列类型
• descriptions_scores: 整数序列类型，使用int64
• top_description: 字符串类型

数据集分割

• train: 训练集
  • 数据量: 54208291.0 字节
  • 样本数: 655

数据集大小

• 下载大小: 14813433 字节
• 数据集总大小: 54208291.0 字节

配置

• default
  • 数据文件路径: data/train-*