---
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
dataset_info:
  features:
  - name: image
    dtype: image
  - name: class
    dtype: string
  - name: id
    dtype: string
  splits:
  - name: train
    num_bytes: 3879773546902
    num_examples: 13153480
  download_size: 3876745904054
  dataset_size: 3879773546902
---

NOTE: I have recaptioned all images [here](https://huggingface.co/datasets/gmongaras/Imagenet21K_Recaption)

This dataset is the entire 21K ImageNet dataset with about 13 million examples and about 19 thousand classes as strings 
(for some reason it only had ~19K classes instead of 21K).

The images are in PNG format. They can be decoded like in the following example
```
import io
from PIL import Image
Image.open(io.BytesIO(row["image"]))
```
where row["image"] are the raw image bytes.