edbeeching/gia-dataset-tokenized-debug

---
dataset_info:
config_name: atari-alien
features:
- name: input_types
sequence: int64
- name: input_ids
sequence: int32
- name: patches
sequence:
sequence:
sequence:
sequence: uint8
- name: loss_mask
sequence: bool
- name: patch_positions
sequence:
sequence:
sequence: float64
- name: local_positions
sequence: int64
- name: attention_mask
sequence: bool
splits:
- name: test
num_bytes: 442153668
num_examples: 335
- name: train
num_bytes: 381345596
num_examples: 289
download_size: 66896005
dataset_size: 823499264
configs:
- config_name: atari-alien
data_files:
- split: test
path: atari-alien/test-*
- split: train
path: atari-alien/train-*
---
# Dataset Card for "gia-dataset-tokenized-debug"

[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

数据集信息

• 配置名称: atari-alien

特征

• input_types: 序列类型为 int64
• input_ids: 序列类型为 int32
• patches: 序列类型为多层嵌套序列，最内层为 uint8
• loss_mask: 序列类型为 bool
• patch_positions: 序列类型为多层嵌套序列，最内层为 float64
• local_positions: 序列类型为 int64
• attention_mask: 序列类型为 bool

数据分割

• 测试集:
  • 名称: test
  • 字节数: 442153668
  • 样本数: 335
• 训练集:
  • 名称: train
  • 字节数: 381345596
  • 样本数: 289

数据大小

• 下载大小: 66896005 字节
• 数据集大小: 823499264 字节

配置文件

• 配置名称: atari-alien
• 数据文件:
  • 测试集路径: atari-alien/test-*
  • 训练集路径: atari-alien/train-*