Datasets:

McCheng
/

CaltechFN_cleaned

image imagewidth (px) 5 129	label class label 10 classes	filename stringlengths 5 10
	99	1.png
	55	8.png
	77	11.png
	66	12.png
	66	13.png
	44	14.png
	55	15.png
	55	16.png
	66	18.png
	88	19.png
	44	21.png
	22	22.png
	88	23.png
	22	25.png
	55	26.png
	22	27.png
	11	30.png
	22	31.png
	11	32.png
	11	33.png
	99	34.png
	55	35.png
	66	38.png
	44	39.png
	88	40.png
	88	42.png
	33	43.png
	22	44.png
	22	46.png
	77	48.png
	88	53.png
	88	55.png
	44	56.png
	22	57.png
	55	58.png
	11	59.png
	99	60.png
	77	61.png
	33	62.png
	77	67.png
	88	68.png
	77	69.png
	11	70.png
	66	71.png
	55	75.png
	88	81.png
	22	84.png
	99	86.png
	55	87.png
	99	88.png
	55	91.png
	99	92.png
	44	93.png
	99	94.png
	11	98.png
	44	100.png
	55	101.png
	22	104.png
	22	106.png
	99	111.png
	99	112.png
	99	116.png
	55	117.png
	88	118.png
	55	119.png
	00	122.png
	11	125.png
	33	130.png
	88	132.png
	22	133.png
	33	135.png
	77	136.png
	55	137.png
	22	139.png
	55	141.png
	44	143.png
	88	144.png
	88	145.png
	77	147.png
	88	152.png
	11	155.png
	55	160.png
	77	165.png
	44	168.png
	00	170.png
	33	173.png
	44	175.png
	00	176.png
	22	177.png
	77	180.png
	55	182.png
	11	183.png
	99	184.png
	55	187.png
	88	188.png
	55	189.png
	66	190.png
	99	192.png
	99	194.png
	00	196.png

End of preview. Expand in Data Studio

Caltech Football Numbers (CaltechFN) Cleaned

This version has been meticulously cleaned to improve data quality for model training. A Vision Large Language Model (google/gemma-3-4b-it) was used to verify the label of each image. Images that were corrupted, unclear, or had incorrect labels were removed.

The final filtered dataset contains:

Training set: 100,442 images
Test set: 25,077 images

Downloads last month: 22

Size of downloaded dataset files:

320 MB

Size of the auto-converted Parquet files:

320 MB

Number of rows:

125,519