---
language: 
- en
tags:
- dataset-compression
- dataset-pruning
- ipc-50
---

# Dataset used for paper -> "[Rethinking Dataset Compression: Shifting Focus From Labels to Images](https://arxiv.org/abs/2502.06434)"
Dataset created according to the paper [Deep Learning on a Data Diet: Finding Important Examples Early in Training](https://arxiv.org/abs/2107.07075).

## Basic Usage

```python
from datasets import load_dataset
dataset = load_dataset("he-yang/2025-rethinkdc-imagenet-el2n-ipc-50")
```

For more information, please refer to the [Rethinking-Dataset-Compression](https://github.com/ArmandXiao/Rethinking-Dataset-Compression)