Datasets:

Modalities:
Text
Formats:
parquet
ArXiv:
Libraries:
Datasets
Dask
License:
Research-EAI commited on
Commit
353370d
Β·
verified Β·
1 Parent(s): 5e1335b

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +2 -2
README.md CHANGED
@@ -4,13 +4,13 @@ license: apache-2.0
4
 
5
  # πŸ’» EAI-Taxonomy Code w/ DCLM
6
 
7
- A high-quality code dataset curated from web data using taxonomy-based filtering, containing **564 billion tokens** of web code content.
8
 
9
  ## 🎯 Dataset Overview
10
 
11
  This dataset is part of the [**Essential-Web**](https://huggingface.co/datasets/EssentialAI/essential-web) project, which introduces a new paradigm for dataset curation using expressive metadata and simple semantic filters. Unlike traditional code datasets that require complex domain-specific pipelines, our approach leverages a 12-category taxonomy to efficiently identify and extract high-quality code data.
12
 
13
- **πŸ’‘ EAI-Taxonomy Code w/ DCLM** (564B tokens): Documents targeting code documentation that exhibit intermediate to advanced reasoning, combined with the DCLM classifier to filter for instruction-dense documents. Also includes mathematics content (`51 - Mathematics`) to match the scope of existing code datasets.
14
 
15
  ## πŸ† Performance
16
 
 
4
 
5
  # πŸ’» EAI-Taxonomy Code w/ DCLM
6
 
7
+ A high-quality code dataset curated from web data using taxonomy-based filtering, containing **100 billion tokens** of web code content.
8
 
9
  ## 🎯 Dataset Overview
10
 
11
  This dataset is part of the [**Essential-Web**](https://huggingface.co/datasets/EssentialAI/essential-web) project, which introduces a new paradigm for dataset curation using expressive metadata and simple semantic filters. Unlike traditional code datasets that require complex domain-specific pipelines, our approach leverages a 12-category taxonomy to efficiently identify and extract high-quality code data.
12
 
13
+ **πŸ’‘ EAI-Taxonomy Code w/ DCLM** (100B tokens): Documents targeting code documentation that exhibit intermediate to advanced reasoning, combined with the DCLM classifier to filter for instruction-dense documents. Also includes mathematics content (`51 - Mathematics`) to match the scope of existing code datasets.
14
 
15
  ## πŸ† Performance
16