Datasets:

gmongaras
/

Imagenet21K

Modalities:

Image

Text

Formats:

Size:

Libraries:

Dataset card Data Studio Files Files and versions

xet

Community

Dataset Viewer

Auto-converted to Parquet Duplicate

Split (1)

train · 13.2M rows

image imagewidth (px) 45 4.37k	class stringclasses 16 values	id stringlengths 11 16
	organism, being	n00004475_54295
	organism, being	n00004475_41022
	organism, being	n00004475_6590
	organism, being	n00004475_35466
	organism, being	n00004475_39382
	organism, being	n00004475_32312
	organism, being	n00004475_15899
	organism, being	n00004475_42770
	benthos	n00005787_8519
	benthos	n00005787_8090
	benthos	n00005787_11666
	benthos	n00005787_3296
	benthos	n00005787_2925
	benthos	n00005787_1318
	benthos	n00005787_11220
	benthos	n00005787_9674
	benthos	n00005787_14527
	benthos	n00005787_1257
	benthos	n00005787_909
	benthos	n00005787_5515
	benthos	n00005787_7386
	benthos	n00005787_951
	benthos	n00005787_8784
	benthos	n00005787_14354
	benthos	n00005787_3498
	benthos	n00005787_8389
	benthos	n00005787_5585
	benthos	n00005787_4337
	benthos	n00005787_12631
	benthos	n00005787_382
	benthos	n00005787_7166
	benthos	n00005787_2478
	benthos	n00005787_14921
	benthos	n00005787_3833
	benthos	n00005787_13420
	benthos	n00005787_10877
	benthos	n00005787_9333
	benthos	n00005787_13774
	benthos	n00005787_13218
	benthos	n00005787_13083
	benthos	n00005787_5851
	benthos	n00005787_2960
	benthos	n00005787_4556
	benthos	n00005787_7864
	benthos	n00005787_7673
	benthos	n00005787_13496
	benthos	n00005787_10760
	benthos	n00005787_14503
	benthos	n00005787_5367
	benthos	n00005787_2237
	benthos	n00005787_2551
	benthos	n00005787_10011
	benthos	n00005787_3979
	benthos	n00005787_3943
	benthos	n00005787_7567
	benthos	n00005787_5267
	benthos	n00005787_3384
	benthos	n00005787_885
	benthos	n00005787_3750
	benthos	n00005787_13413
	benthos	n00005787_12823
	benthos	n00005787_7298
	benthos	n00005787_6372
	benthos	n00005787_5865
	benthos	n00005787_8504
	benthos	n00005787_1739
	benthos	n00005787_7370
	benthos	n00005787_6830
	benthos	n00005787_12408
	benthos	n00005787_457
	benthos	n00005787_7801
	benthos	n00005787_14044
	benthos	n00005787_3874
	benthos	n00005787_236
	benthos	n00005787_7287
	benthos	n00005787_2404
	benthos	n00005787_14237
	benthos	n00005787_10703
	benthos	n00005787_2718
	benthos	n00005787_4101
	benthos	n00005787_11955
	benthos	n00005787_7359
	benthos	n00005787_6652
	benthos	n00005787_4506
	benthos	n00005787_4239
	benthos	n00005787_5070
	benthos	n00005787_5398
	benthos	n00005787_13
	benthos	n00005787_8061
	benthos	n00005787_2553
	benthos	n00005787_8735
	benthos	n00005787_10944
	benthos	n00005787_4585
	benthos	n00005787_9885
	benthos	n00005787_3169
	benthos	n00005787_1408
	benthos	n00005787_13108
	benthos	n00005787_13639
	benthos	n00005787_15093
	benthos	n00005787_12029

End of preview. Expand in Data Studio

NOTE: I have recaptioned all images here

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.

Downloads last month: 10,299

Size of downloaded dataset files:

3.88 TB

Size of the auto-converted Parquet files:

3.88 TB

Number of rows:

13,153,480

Space using gmongaras/Imagenet21K 1

Collection including gmongaras/Imagenet21K

datasets

Collection

11 items • Updated Apr 23, 2025