#858

/capstor/store/cscs/swissai/infra01/vision-datasets/raw/stage2/mvp-lab__LLaVA-OneVision-1.5-Mid-Training-85M

kind	parquet
status	active
samples	234,597,376
counted via	parquet footer
size	46.7 TB
files	12,176
first seen	2026-07-22 14:13
last seen	2026-07-22 14:13
registered	2026-07-22 14:13

# dataset #858
path: /capstor/store/cscs/swissai/infra01/vision-datasets/raw/stage2/mvp-lab__LLaVA-OneVision-1.5-Mid-Training-85M
kind: parquet | status: active
samples: 234,597,376 (parquet footer) | size: 46.7 TB | files: 12,176

fields:
- id: text
- image: image
- caption: text

tree (truncated):
```
.gitattributes
README.md
coyo/EN/part00/train-00000-of-00063.parquet
coyo/EN/part00/train-00001-of-00063.parquet
coyo/EN/part00/train-00002-of-00063.parquet
coyo/EN/part00/train-00003-of-00063.parquet
coyo/EN/part00/train-00004-of-00063.parquet
coyo/EN/part00/train-00005-of-00063.parquet
coyo/EN/part00/train-00006-of-00063.parquet
coyo/EN/part00/train-00007-of-00063.parquet
coyo/EN/part00/train-00008-of-00063.parquet
coyo/EN/part00/train-00009-of-00063.parquet
coyo/EN/part00/train-00010-of-00063.parquet
coyo/EN/part00/train-00011-of-00063.parquet
coyo/EN/part00/train-00012-of-00063.parquet
coyo/EN/part00/train-00013-of-00063.parquet
coyo/EN/part00/train-00014-of-00063.parquet
coyo/EN/part00/train-00015-of-00063.parquet
coyo/EN/part00/train-00016-of-00063.parquet
coyo/EN/part00/train-00017-of-00063.parquet
coyo/EN/part00/train-00018-of-00063.parquet
coyo/EN/part00/train-00019-of-00063.parquet
coyo/EN/part00/train-00020-of-00063.parquet
coyo/EN/part00/train-00021-of-00063.parquet
coyo/EN/part00/train-00022-of-00063.parquet
... (+12,151 more)
```

samples


The image features a bold, three-dimensional orange text sculpture spelling "DSAAC" against a plain white background. The letters are thick, rounded, and slightly tilted, giving a playful, modern feel. They cast soft shadows on the white surface below, indicating a light source from above, which enhances their depth. The smooth, matte texture of the orange letters contrasts with the clean, seamless white backdrop. The overall composition is simple, with the text centered and occupying most of the frame, creating a focal point that draws immediate attention. There’s no additional context or elements, making the focus solely on the typography. The bright orange color conveys energy and warmth, while the minimalistic setting keeps the design uncluttered and impactful.
The image is a line graph titled "Europe Industrial Production" with the Alhambra Investments logo at the top right. It measures industrial production (mining, manufacturing, utilities) from 1990 to 2020, indexed to 2015=100 and seasonally adjusted. A green line represents Italy, showing a peak around 2007-2008 near 130, then a sharp drop to ~90 by 2009, partial recovery, another decline post-2012, and a final steep drop to ~70 by 2020. Red text questions "Europe is booming?" and states "Italy never recovered from neither 2012 nor 2008, and now...". The x-axis shows years; y-axis ranges 70-130. The graph, with a white background, uses green for data and red for annotations, highlighting Italy's long-term industrial decline, relevant to economic analysis for investors or economists.
A digital map on a screen displays a green flight path from Chicago to Manchester, with various cities like Calgary, Denver, and Reykjavik labeled in white text. The map shows parts of North America and Europe, with landmasses in green/brown and oceans in dark blue. Below the map, text reads "Time to Destination 8:06" and "Local Time at Present Position 1:03 pm." The screen has a golden frame, and small blue labels mark underwater features like "Charlie-Gibbs Fracture Zone." The perspective is a zoomed-in view of the Northern Hemisphere, suggesting a flight tracking interface, likely for an audience interested in travel or navigation, with clear spatial arrangement of cities along the route.
The image features a dark, eerie scene with a desolate, industrial background, including blurred metal structures and a dimly lit corridor with warm orange-red hues. Prominently displayed in bright yellow text is the quote: "Death has been here and where death has been no light shall ever be." Below it, in smaller yellow text, is the attribution "~ Courtney Summers," and at the bottom center, "WISEFAMOUSQUOTES.COM" is written in white capital letters. The overall atmosphere is somber and foreboding, enhanced by the low lighting and shadowy, decaying environment. The text is the focal point, centered against the dark backdrop, creating a stark contrast that emphasizes the quote's grim message about permanence of darkness after death. The scene suggests an abandoned or post-apocalyptic setting, contributing to the heavy, melancholic mood.
The image depicts a busy indoor science fair or academic exhibition, likely at a high school or university. Multiple blue and black poster boards with papers, graphs, and photos are set up on tables, each displaying student projects. People of various ages, including students and adults, move around, some stopping to read posters while others converse. In the foreground, two young women chat: one wears a blue shirt with yellow text on the back and has short blonde hair, while the other wears a brown sweater and holds a blue folder, with a name tag pinned to her chest. The background shows large windows letting in natural daylight, brightening the spacious hall with white walls and wooden accents. The atmosphere is lively and engaged, with attendees focused on learning and discussing the displayed work, suggesting a community event celebrating education and research.
The image features two small, oval-shaped objects with a brown, textured base and irregular white patches resembling mineral formations, likely geode slices or polished stones. They are positioned side by side against a plain white background, centered in the frame with minimal spacing between them. The lighting is even, highlighting the rough, matte texture of the brown edges and the slightly smoother, crystalline appearance of the white areas. There are no people, text, or other objects present, creating a simple, focused composition. The color palette is neutral, with earthy browns and off-whites, conveying a natural, organic feel. The objects are similar in size and shape, suggesting they might be a matched pair, perhaps intended for use as decorative items or jewelry components. The lack of context or environment keeps the focus on their physical attributes—elliptical form, contrasting color patterns, and tactile surface—making the image feel like a静物 study of natural materials.
The image features a clear glass filled with vibrant red pomegranate juice on a white surface, alongside a whole pomegranate, a pile of its ruby-red arils, and a small bowl of arils with some juice splattered nearby. Above these elements, red text reads "Pomegranate Juice 'Significantly Improves' Memory," highlighting a potential health claim. The setting is minimal, with a white background that emphasizes the rich red hues of the fruit and juice, creating a clean, fresh atmosphere. The lighting is bright and even, casting soft shadows that enhance the textures of the arils and the smooth surface of the juice. The composition places the glass of juice on the left, the whole pomegranate in the center, and the bowl of arils on the right, balancing the scene. The splattered juice adds a dynamic touch, suggesting recent preparation. This image likely targets health-conscious individuals interested in natural remedies or nutritional benefits, using vivid colors to appeal visually and text to convey a specific message about cognitive improvement.
A vintage silver drag racing car with the number 7398 and "LAFAYETTE MOTORSPORTS RACING TEAM" written on its windshield is captured mid-launch, its front wheels lifted off the drag strip, showcasing the powerful acceleration typical of drag racing events. The car features a large grille, round headlights, and wide rear tires, with sponsor decals visible on its side. In the background, two men stand: one in a red "BSD" shirt to the left, and another in a yellow shirt pushing a bike to the right, suggesting a busy pit or trackside area. The scene is outdoors on a sunny day, with the drag strip's smooth, dark asphalt contrasting with the car's metallic finish. The car’s raised front end and the men’s casual stances convey excitement and activity, typical of a motorsports event, where speed and precision are key. The overall mood is dynamic and energetic, highlighting the thrill of drag racing.
The image is a bar chart titled "Leistung im internationalen Datenschutztest" (Performance in the International Data Protection Test) with a German flag icon and "Deutschland" (Germany) below. Four categories are shown: Gesamtergebnis des Datenschutztests (Overall Data Protection Test Result), Digitale Gewohnheiten (Digital Habits), Digitales Datenschutzwissen (Digital Data Protection Knowledge), and Online-Riskomanagement (Online Risk Management). Each category has two bars: a blue one labeled "Deutschland" (Germany) and a light gray one labeled "Weltweit" (Worldwide). Germany scores 71.2 vs. worldwide 65.2 (+6), 83.2 vs. 77.1 (+6.1), 78 vs. 72.2 (+5.8), and 90.2 vs. 84.2 (+6). The chart uses a clean, minimal design with green "+" indicators for Germany's positive differences, set against a white background. The perspective is a standard top-down view, making data comparison straightforward, targeting a general audience interested in data protection performance metrics.
The image shows an indoor setting with a patterned maroon and gold wall in the background. To the right, a large vertical banner displays "YOUTH FOR DEMOCRACY" in bold red and black text at the top, with Arabic script below that translates to "يدعم الشباب الديمقراطية" (Youth support democracy). Below the Arabic text are small icons of a person and speech bubbles, suggesting community or dialogue. To the left of the banner stands an easel with a whiteboard that has faint, colorful scribbles or drawings, though the details are indistinct. The bottom left corner of the image has small logos of various organizations, indicating possible sponsors. The lighting is warm, casting soft shadows, creating a formal yet engaging atmosphere, likely for a youth-focused event or workshop centered on democratic engagement. The spatial arrangement places the banner as the focal point, with the easel positioned slightly behind it, balancing the composition. The color palette includes deep reds, blues, and golds, contributing to a sense of purpose and vibrancy.

#	id	caption
1	coyo/part_01/coyo700m_00_id_10000029.jpg	The image features a bold, three-dimensional orange text sculpture spelling "DSAAC" against a plain white background. The letters are thick, rounded, and slightly tilted, giving a playful, modern feel. They cast soft shadows on the white surface below, indicating a light source from above, which enhances their depth. The smooth, matte texture of the orange letters contrasts with the clean, seamless white backdrop. The overall composition is simple, with the text centered and occupying most of the frame, creating a focal point that draws immediate attention. There’s no additional context or elements, making the focus solely on the typography. The bright orange color conveys energy and warmth, while the minimalistic setting keeps the design uncluttered and impactful.
2	coyo/part_01/coyo700m_00_id_10000046.jpg	The image is a line graph titled "Europe Industrial Production" with the Alhambra Investments logo at the top right. It measures industrial production (mining, manufacturing, utilities) from 1990 to 2020, indexed to 2015=100 and seasonally adjusted. A green line represents Italy, showing a peak around 2007-2008 near 130, then a sharp drop to ~90 by 2009, partial recovery, another decline post-2012, and a final steep drop to ~70 by 2020. Red text questions "Europe is booming?" and states "Italy never recovered from neither 2012 nor 2008, and now...". The x-axis shows years; y-axis ranges 70-130. The graph, with a white background, uses green for data and red for annotations, highlighting Italy's long-term industrial decline, relevant to economic analysis for investors or economists.
3	coyo/part_01/coyo700m_00_id_10000250.jpg	A digital map on a screen displays a green flight path from Chicago to Manchester, with various cities like Calgary, Denver, and Reykjavik labeled in white text. The map shows parts of North America and Europe, with landmasses in green/brown and oceans in dark blue. Below the map, text reads "Time to Destination 8:06" and "Local Time at Present Position 1:03 pm." The screen has a golden frame, and small blue labels mark underwater features like "Charlie-Gibbs Fracture Zone." The perspective is a zoomed-in view of the Northern Hemisphere, suggesting a flight tracking interface, likely for an audience interested in travel or navigation, with clear spatial arrangement of cities along the route.
4	coyo/part_01/coyo700m_00_id_1000016.jpg	The image features a dark, eerie scene with a desolate, industrial background, including blurred metal structures and a dimly lit corridor with warm orange-red hues. Prominently displayed in bright yellow text is the quote: "Death has been here and where death has been no light shall ever be." Below it, in smaller yellow text, is the attribution "~ Courtney Summers," and at the bottom center, "WISEFAMOUSQUOTES.COM" is written in white capital letters. The overall atmosphere is somber and foreboding, enhanced by the low lighting and shadowy, decaying environment. The text is the focal point, centered against the dark backdrop, creating a stark contrast that emphasizes the quote's grim message about permanence of darkness after death. The scene suggests an abandoned or post-apocalyptic setting, contributing to the heavy, melancholic mood.
5	coyo/part_01/coyo700m_00_id_10000103.jpg	The image depicts a busy indoor science fair or academic exhibition, likely at a high school or university. Multiple blue and black poster boards with papers, graphs, and photos are set up on tables, each displaying student projects. People of various ages, including students and adults, move around, some stopping to read posters while others converse. In the foreground, two young women chat: one wears a blue shirt with yellow text on the back and has short blonde hair, while the other wears a brown sweater and holds a blue folder, with a name tag pinned to her chest. The background shows large windows letting in natural daylight, brightening the spacious hall with white walls and wooden accents. The atmosphere is lively and engaged, with attendees focused on learning and discussing the displayed work, suggesting a community event celebrating education and research.
6	coyo/part_01/coyo700m_00_id_10000143.jpg	The image features two small, oval-shaped objects with a brown, textured base and irregular white patches resembling mineral formations, likely geode slices or polished stones. They are positioned side by side against a plain white background, centered in the frame with minimal spacing between them. The lighting is even, highlighting the rough, matte texture of the brown edges and the slightly smoother, crystalline appearance of the white areas. There are no people, text, or other objects present, creating a simple, focused composition. The color palette is neutral, with earthy browns and off-whites, conveying a natural, organic feel. The objects are similar in size and shape, suggesting they might be a matched pair, perhaps intended for use as decorative items or jewelry components. The lack of context or environment keeps the focus on their physical attributes—elliptical form, contrasting color patterns, and tactile surface—making the image feel like a静物 study of natural materials.
7	coyo/part_01/coyo700m_00_id_10000042.jpg	The image features a clear glass filled with vibrant red pomegranate juice on a white surface, alongside a whole pomegranate, a pile of its ruby-red arils, and a small bowl of arils with some juice splattered nearby. Above these elements, red text reads "Pomegranate Juice 'Significantly Improves' Memory," highlighting a potential health claim. The setting is minimal, with a white background that emphasizes the rich red hues of the fruit and juice, creating a clean, fresh atmosphere. The lighting is bright and even, casting soft shadows that enhance the textures of the arils and the smooth surface of the juice. The composition places the glass of juice on the left, the whole pomegranate in the center, and the bowl of arils on the right, balancing the scene. The splattered juice adds a dynamic touch, suggesting recent preparation. This image likely targets health-conscious individuals interested in natural remedies or nutritional benefits, using vivid colors to appeal visually and text to convey a specific message about cognitive improvement.
8	coyo/part_01/coyo700m_00_id_10000435.jpg	A vintage silver drag racing car with the number 7398 and "LAFAYETTE MOTORSPORTS RACING TEAM" written on its windshield is captured mid-launch, its front wheels lifted off the drag strip, showcasing the powerful acceleration typical of drag racing events. The car features a large grille, round headlights, and wide rear tires, with sponsor decals visible on its side. In the background, two men stand: one in a red "BSD" shirt to the left, and another in a yellow shirt pushing a bike to the right, suggesting a busy pit or trackside area. The scene is outdoors on a sunny day, with the drag strip's smooth, dark asphalt contrasting with the car's metallic finish. The car’s raised front end and the men’s casual stances convey excitement and activity, typical of a motorsports event, where speed and precision are key. The overall mood is dynamic and energetic, highlighting the thrill of drag racing.
9	coyo/part_01/coyo700m_00_id_10000444.jpg	The image is a bar chart titled "Leistung im internationalen Datenschutztest" (Performance in the International Data Protection Test) with a German flag icon and "Deutschland" (Germany) below. Four categories are shown: Gesamtergebnis des Datenschutztests (Overall Data Protection Test Result), Digitale Gewohnheiten (Digital Habits), Digitales Datenschutzwissen (Digital Data Protection Knowledge), and Online-Riskomanagement (Online Risk Management). Each category has two bars: a blue one labeled "Deutschland" (Germany) and a light gray one labeled "Weltweit" (Worldwide). Germany scores 71.2 vs. worldwide 65.2 (+6), 83.2 vs. 77.1 (+6.1), 78 vs. 72.2 (+5.8), and 90.2 vs. 84.2 (+6). The chart uses a clean, minimal design with green "+" indicators for Germany's positive differences, set against a white background. The perspective is a standard top-down view, making data comparison straightforward, targeting a general audience interested in data protection performance metrics.
10	coyo/part_01/coyo700m_00_id_1000042.jpg	The image shows an indoor setting with a patterned maroon and gold wall in the background. To the right, a large vertical banner displays "YOUTH FOR DEMOCRACY" in bold red and black text at the top, with Arabic script below that translates to "يدعم الشباب الديمقراطية" (Youth support democracy). Below the Arabic text are small icons of a person and speech bubbles, suggesting community or dialogue. To the left of the banner stands an easel with a whiteboard that has faint, colorful scribbles or drawings, though the details are indistinct. The bottom left corner of the image has small logos of various organizations, indicating possible sponsors. The lighting is warm, casting soft shadows, creating a formal yet engaging atmosphere, likely for a youth-focused event or workshop centered on democratic engagement. The spatial arrangement places the banner as the focal point, with the easel positioned slightly behind it, balancing the composition. The color palette includes deep reds, blues, and golds, contributing to a sense of purpose and vibrancy.