---
configs:
- config_name: default
  data_files:
  - split: train
    path:
    - "*.jsonl.zst"
  default: true
annotations_creators:
- found
language:
- ru
- en
- kk
- uk
- be
- multilingual
license: cc0-1.0
multilinguality:
- multilingual
pretty_name: cdnpdf Educational Materials Dataset (Part 1)
size_categories:
- 100K<n<1M
source_datasets:
- original
task_categories:
- text-classification
- text-retrieval
---
# Dataset Card for cdnpdf Educational Materials (Part 1)
## Dataset Summary
This dataset contains metadata and original files for 101,022 educational presentations from the [cdnpdf.com](https://cdnpdf.com) platform, which provides free access to books, documents, magazines and presentations. This collection focuses exclusively on presentations and includes archives with IDs from 00 to 24. The dataset includes information such as presentation titles, descriptions, URLs, download URLs, and file paths. This is Part 1 of the cdnpdf dataset, containing presentations with archive IDs from 00 to 24. All PPT files have been converted to PPTX format for better compatibility and reduced file size.

Part 2 of this dataset is available at https://huggingface.co/datasets/nyuuzyou/cdnpdf-presentations-part2

## Languages
The dataset is primarily in Russian, with some presentations in other languages such as English, Kazakh, Ukrainian, and Belarusian.

## Dataset Structure
### Data Files
- Metadata: Stored in JSON Lines format
- Original files: PPTX presentations (PPT files were converted to PPTX for better compatibility and smaller size)

### Data Fields
- `url`: URL of the presentation page on cdnpdf.com
- `title`: Title of the educational presentation
- `description`: Description of the presentation
- `download_url`: URL for downloading the presentation
- `filepath`: Local filepath of the downloaded PPTX file

## Additional Information
### License
This dataset is dedicated to the public domain under the Creative Commons Zero (CC0) license, allowing for any use, modification, or distribution without attribution.

### Dataset Curators
- [nyuuzyou](https://ducks.party)