morten-j/medhie-tokenized-dataset-custom-bertwordpiece

---
dataset_info:
features:
- name: input_ids
sequence: int32
- name: token_type_ids
sequence: int8
- name: attention_mask
sequence: int8
- name: word_ids
sequence: int64
- name: labels
sequence: int64
splits:
- name: train
num_bytes: 42113819652
num_examples: 14830301
- name: validation
num_bytes: 4836576832
num_examples: 1703172
download_size: 21646576639
dataset_size: 46950396484
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
---

原始信息汇总

数据集信息

特征

• input_ids: 序列类型为 int32
• token_type_ids: 序列类型为 int8
• attention_mask: 序列类型为 int8
• word_ids: 序列类型为 int64
• labels: 序列类型为 int64

数据分割

• 训练集 (train):
  • 字节数: 42113819652
  • 样本数: 14830301
• 验证集 (validation):
  • 字节数: 4836576832
  • 样本数: 1703172

数据大小

• 下载大小: 21646576639 字节
• 数据集大小: 46950396484 字节

配置

• 默认配置 (default):
  • 训练集文件路径: data/train-*
  • 验证集文件路径: data/validation-*