Datasets:

leobottaro
/

handwritten-digits-pack

Tasks:

Image Classification

Modalities:

Formats:

Languages:

Size:

Tags:

Libraries:

License:

Dataset card Data Studio Files Files and versions

xet

Community

Dataset Viewer

Auto-converted to Parquet Duplicate

Split (2)

train · 503k rows

Search is not available for this dataset

image imagewidth (px) 28 28	label int32 0 9
	1
	2
	0
	5
	9
	6
	2
	7
	9
	2
	6
	3
	4
	8
	7
	2
	8
	9
	7
	5
	1
	4
	0
	8
	9
	6
	6
	4
	5
	8
	3
	0
	9
	3
	3
	5
	7
	5
	8
	4
	6
	2
	3
	6
	7
	1
	1
	5
	2
	7
	1
	6
	1
	4
	1
	1
	7
	7
	1
	6
	8
	9
	2
	0
	4
	3
	0
	5
	2
	9
	4
	1
	5
	2
	1
	5
	4
	0
	6
	3
	6
	1
	9
	3
	2
	4
	5
	1
	3
	3
	5
	9
	9
	0
	8
	9
	8
	4
	6
	2

End of preview. Expand in Data Studio

Handwritten Digits Pack

Unified 28×28 grayscale handwritten digit dataset for training classification models (digits 0–9). White background, dark digits, float32 [0, 1].

Composition

Merged from 3 public datasets:

Dataset	Train	Validation	Source
EMNIST Digits	280,000	—	NIST
HWD-V1	136,920	15,218	Kaggle
Handwritten Digits	86,180	21,540	HF
Total	503,100	36,758

Format

train.parquet — 503,100 rows (191.8 MB)
validation.parquet — 36,758 rows (11.0 MB)
image column: PNG-encoded 28×28 grayscale (struct with bytes + path fields)
label column: int32, digits 0–9

Loading

from datasets import load_dataset
ds = load_dataset("leobottaro/handwritten-digits-pack")

Preprocessing Applied

EMNIST: IDX binary → transpose (fix orientation) → invert (white bg, dark digit) → float32
HWD-V1: 52×52 PNG → LANCZOS resize 28×28 → grayscale → float32. Includes both Standard and Edge Cases
Handwritten Digits: RGBA PNG → composite on white background (preserves dark-on-transparent) → grayscale → float32

Credits

Cohen, G., Afshar, S., Tapson, J., & van Schaik, A. (2017). EMNIST: an extension of MNIST to handwritten letters.
Metricas Ecuador — HWD-V1 on Kaggle
nguyenminh4099 — Handwritten Digits on Hugging Face

Downloads last month: -

image imagewidth (px) 28 28	label int32 0 9
	1
	2
	0
	5
	9
	6
	2
	7
	9
	2
	6
	3
	4
	8
	7
	2
	8
	9
	7
	5
	1
	4
	0
	8
	9
	6
	6
	4
	5
	8
	3
	0
	9
	3
	3
	5
	7
	5
	8
	4
	6
	2
	3
	6
	7
	1
	1
	5
	2
	7
	1
	6
	1
	4
	1
	1
	7
	7
	1
	6
	8
	9
	2
	0
	4
	3
	0
	5
	2
	9
	4
	1
	5
	2
	1
	5
	4
	0
	6
	3
	6
	1
	9
	3
	2
	4
	5
	1
	3
	3
	5
	9
	9
	0
	8
	9
	8
	4
	6
	2

image imagewidth (px) 28 28	label int32 0 9
	1
	2
	0
	5
	9
	6
	2
	7
	9
	2
	6
	3
	4
	8
	7
	2
	8
	9
	7
	5
	1
	4
	0
	8
	9
	6
	6
	4
	5
	8
	3
	0
	9
	3
	3
	5
	7
	5
	8
	4
	6
	2
	3
	6
	7
	1
	1
	5
	2
	7
	1
	6
	1
	4
	1
	1
	7
	7
	1
	6
	8
	9
	2
	0
	4
	3
	0
	5
	2
	9
	4
	1
	5
	2
	1
	5
	4
	0
	6
	3
	6
	1
	9
	3
	2
	4
	5
	1
	3
	3
	5
	9
	9
	0
	8
	9
	8
	4
	6
	2

image imagewidth (px) 28 28	label int32 0 9
	1
	2
	0
	5
	9
	6
	2
	7
	9
	2
	6
	3
	4
	8
	7
	2
	8
	9
	7
	5
	1
	4
	0
	8
	9
	6
	6
	4
	5
	8
	3
	0
	9
	3
	3
	5
	7
	5
	8
	4
	6
	2
	3
	6
	7
	1
	1
	5
	2
	7
	1
	6
	1
	4
	1
	1
	7
	7
	1
	6
	8
	9
	2
	0
	4
	3
	0
	5
	2
	9
	4
	1
	5
	2
	1
	5
	4
	0
	6
	3
	6
	1
	9
	3
	2
	4
	5
	1
	3
	3
	5
	9
	9
	0
	8
	9
	8
	4
	6
	2