Datasets:

Modalities:
Tabular
Text
Formats:
parquet
ArXiv:
Libraries:
Datasets
Dask
License:
File size: 652 Bytes
2190446
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
---
license: odc-by
dataset_info:
  features:
  - name: url
    dtype: string
  - name: tag
    dtype: string
  - name: text
    dtype: string
  - name: file_path
    dtype: string
  - name: dump
    dtype: string
  - name: file_size_in_byte
    dtype: int64
  - name: line_count
    dtype: int64
  splits:
  - name: train
    num_bytes: 18159796472
    num_examples: 5241900
  download_size: 9949701917
  dataset_size: 18159796472
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---

This data corpus is a math corpus recalled from [Fineweb](https://huggingface.co/datasets/HuggingFaceFW/fineweb) using fastText.