amitness/logits-mt-ar-128

---
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*
dataset_info:
features:
- name: input_ids
sequence: int32
- name: token_type_ids
sequence: int8
- name: attention_mask
sequence: int8
- name: labels
sequence: int64
- name: teacher_logits
sequence:
sequence: float64
- name: teacher_indices
sequence:
sequence: int64
- name: teacher_mask_indices
sequence: int64
splits:
- name: train
num_bytes: 17264984288
num_examples: 3814324
- name: test
num_bytes: 3047653868
num_examples: 673117
download_size: 2917556992
dataset_size: 20312638156
---
# Dataset Card for "logits-mt-ar-128"

[More Information needed](https://github.com/huggingface/datasets/blob/main/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)

原始信息汇总

数据集概述

配置

• 默认配置
  • 数据文件
    • 训练集: data/train-*
    • 测试集: data/test-*

数据特征

• 输入ID: input_ids，类型为 int32 序列
• 标记类型ID: token_type_ids，类型为 int8 序列
• 注意力掩码: attention_mask，类型为 int8 序列
• 标签: labels，类型为 int64 序列
• 教师对数: teacher_logits，类型为 float64 序列的序列
• 教师索引: teacher_indices，类型为 int64 序列的序列
• 教师掩码索引: teacher_mask_indices，类型为 int64 序列

数据分割

• 训练集
  • 字节数: 17264984288
  • 样本数: 3814324
• 测试集
  • 字节数: 3047653868
  • 样本数: 673117

数据大小

• 下载大小: 2917556992 字节
• 数据集大小: 20312638156 字节