Datasets:

SI-Lab
/

StepCountQA-RL-Traj_11_50_Combined

images images listlengths 1 1	problem stringlengths 36 183	answer stringclasses 40 values
	<image> How many lighter colored Levi apparel are in the image?	17
	<image> How many Spoon are in the image?	30
	<image> How many Total Dominoes are in the image?	24
	<image> How many shoes are in the image?	28
	<image> How many characters in comic are in the image?	22
	<image> How many people are in the image?	22
	<image> How many shoes are in the image?	14
	<image> How many Tennis Balls are in the image?	38
	<image> How many people are in the image?	15
	<image> How many people are in the image?	21
	<image> How many red stripes are in the image?	28
	<image> How many Arrow are in the image?	43
	<image> How many Person are in the image?	25
	<image> How many images in the grid are in the image?	25
	<image> How many small round yellow candies on cake are in the image?	32
	<image> How many white piano keys are in the image?	14
	<image> How many shoes with white edges are in the image?	44
	<image> How many Cells with a date of birth are in the image?	21
	<image> How many chairlifts are in the image?	16
	<image> How many icons are in the image?	37
	<image> How many hot air ballons are in the image?	32
	<image> How many coin are in the image?	20
	<image> How many Butter Knives are in the image?	12
	<image> How many Icons are in the image?	25
	<image> How many bubbles are in the image?	13
	<image> How many Images are in the image?	16
	<image> How many people are in the image?	17
	<image> How many blurred file name are in the image?	27
	<image> How many dancers on stage are in the image?	14
	<image> How many wall sconces are in the image?	12
	<image> How many people are in the image?	18
	<image> How many movies in piles on the left are in the image?	30
	<image> How many Vertical wooden shelf supports are in the image?	32
	<image> How many people are in the image?	11
	<image> How many Christmas Presents are in the image?	13
	<image> How many Leg are in the image?	44
	<image> How many Automobilies are in the image?	40
	<image> How many hangers are in the image?	11
	<image> How many houses are in the image?	11
	<image> How many wine bottles are in the image?	33
	<image> How many cereal boxes are in the image?	11
	<image> How many Stuffed animals are in the image?	22
	<image> How many butterflys are in the image?	13
	<image> How many Books are in the image?	15
	<image> How many stuffed animals are in the image?	24
	<image> How many collage images are in the image?	28
	<image> How many Monopoly spaces are in the image?	40
	<image> How many apps are in the image?	26
	<image> How many glass wall/window are in the image?	16
	<image> How many human being on football field are in the image?	17
	<image> How many glasses are in the image?	17
	<image> How many boats are in the image?	47
	<image> How many car are in the image?	11
	<image> How many automobiles are in the image?	47
	<image> How many white cars are in the image?	13
	<image> How many people are in the image?	14
	<image> How many symbols are in the image?	17
	<image> How many candles are in the image?	11
	<image> How many hole in button the thread goes through are in the image?	12
	<image> How many illustration of hats are in the image?	28
	<image> How many calendar months are in the image?	12
	<image> How many cones are in the image?	21
	<image> How many Red bowling pin stripes. are in the image?	12
	<image> How many people are in the image?	11
	<image> How many pens are in the image?	20
	<image> How many Books are in the image?	11
	<image> How many Seating outside of the restuarant. are in the image?	11
	<image> How many non-meal item are in the image?	15
	<image> How many people are in the image?	38
	<image> How many people are in the image?	12
	<image> How many keyboard buttons are in the image?	45
	<image> How many cars are in the image?	25
	<image> How many fingers are in the image?	40
	<image> How many colored pencils are in the image?	21
	<image> How many player are in the image?	12
	<image> How many heads on black-colored icons are in the image?	15
	<image> How many Items Containing Beef are in the image?	11
	<image> How many book are in the image?	14
	<image> How many people are in the image?	11
	<image> How many plant are in the image?	18
	<image> How many people are in the image?	25
	<image> How many variable names are in the image?	21
	<image> How many face-up cards are in the image?	14
	<image> How many ceiling tiles are in the image?	29
	<image> How many red vehicles are in the image?	16
	<image> How many people are in the image?	42
	<image> How many picture are in the image?	11
	<image> How many Football club are in the image?	29
	<image> How many goggles are in the image?	14
	<image> How many tennis ball are in the image?	19
	<image> How many buttons on television remote control are in the image?	38
	<image> How many pocket clips are in the image?	11
	<image> How many black circles are in the image?	28
	<image> How many words in black font are in the image?	39
	<image> How many Rocks are in the image?	22
	<image> How many English alphabet characters are in the image?	28
	<image> How many cars are in the image?	50
	<image> How many luggages are in the image?	33
	<image> How many VHS Tapes are in the image?	21
	<image> How many people are in the image?	21

End of preview. Expand in Data Studio

StepCountQA-RL-Traj_11_50_Combined

Dataset Description

StepCountQA-RL-Traj_11_50_Combined is a unified visual counting RL training dataset covering object counts from 11 to 50. It merges two complementary 11-50 datasets into a single collection:

Source	Samples	Description
StepCountQA-RL-Traj_11_50	9,508	`<point>` JSON annotation trajectories, converted to numeric
StepCountQA-RL-Traj_11_50_NumericOnly	29,722	Pure numeric answer from SFT trajectories
Combined	39,230	Merged, deduplicated (0 overlap)

All samples pair an original (non-annotated) image with a counting question and a pure numeric answer string.

Dataset Structure

Data Fields

Field	Type	Description
`images`	`Sequence(Image)`	One original image per sample
`problem`	`string`	Counting question with `<image>` tag
`answer`	`string`	Numeric count as string (e.g., `"27"`)

Format Example

{
  "images": [<PIL.Image>],
  "problem": "<image>\nHow many people are in the image?",
  "answer": "39"
}

Dataset Statistics

Total samples: 39,230
Count range: 11 – 50
Unique images: 39,230 (zero duplicates verified)
Total size: ~8.95 GB (4 parquet shards)

Count Distribution

Count	Samples	Count	Samples
11	2,537	31	660
12	3,297	32	676
13	1,863	33	591
14	2,147	34	558
15	2,145	35	621
16	2,011	36	687
17	1,372	37	469
18	1,732	38	484
19	1,235	39	464
20	1,621	40	670
21	1,136	41	425
22	1,072	42	474
23	967	43	359
24	1,182	44	383
25	976	45	424
26	907	46	364
27	753	47	323
28	879	48	432
29	627	49	341
30	940	50	426

Comparison with Related Datasets

Dataset	Samples	Count Range	Answer Format
StepCountQA-RL-Traj_11_50	9,508	11–50	Numeric string
StepCountQA-RL-Traj_11_50_NumericOnly	29,722	11–50	Numeric string
StepCountQA-RL-Traj_11_50_Combined	39,230	11–50	Numeric string

Usage

from datasets import load_dataset

ds = load_dataset("SI-Lab/StepCountQA-RL-Traj_11_50_Combined")
train = ds["train"]
print(train[0])
# {'images': [<PIL.Image>], 'problem': '<image>\nHow many people...', 'answer': '39'}

Dataset Creation

Merged from two sources:

StepCountQA-RL-Traj_11_50 (9,508 rows): Original <point> annotation dataset with numeric answers extracted
StepCountQA-RL-Traj_11_50_NumericOnly (29,722 rows): Converted from 11-50 SFT trajectory dataset

Merge process:

Extracted image path field from both datasets (streaming, 50 rows/batch)
Verified zero overlap (0 duplicate images between A and B)
Streamed A then B in 100-row batches into 4 output shards
Final row count verified: 9,508 + 29,722 = 39,230 ✓

License

Please refer to the source dataset license for usage terms.

Downloads last month: 19