UME-R1
Collection
UME-R1 is a framework designed to endow multimodal embedding models with the flexibility to switch between discriminative and generative embeddings • 5 items • Updated • 9
Datasets:
dataset stringclasses 1
value | dataset_name stringclasses 10
values | qry_text stringlengths 54 1.57k | qry_image stringlengths 0 55 | qry_video listlengths 1 8 | pos_text stringlengths 1 1.73k | pos_image stringlengths 0 51 | pos_video listlengths 1 8 | solution stringclasses 1
value |
|---|---|---|---|---|---|---|---|---|
MMEB-V2-R1 | llavahound_caption_retrieval | Narrate the video's unfolding events in a descriptive manner. | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1066490890/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1066490890/c01_0003.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1066490890/c01_0006.jpeg",
"MMEB-train/llavahound/train_video_and_inst... | In this video, the view is captivated by a green, mossy cliff, sturdy and overwhelming, a spectacle of natural beauty. The moss is a vibrant green, indicating it is well-nourished, thriving from the moisture provided by the trickling waterfall nearby. Water is continuously flowing, painting intriguing patterns as it gr... | [
""
] | <embedding> | ||
MMEB-V2-R1 | llavahound_caption_retrieval | Walk through the video, detailing its key moments and features. | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/17056555/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/17056555/c01_0003.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/17056555/c01_0006.jpeg",
"MMEB-train/llavahound/train_video_and_instructio... | In the uploaded video content, we are presented with a Christmas setting, illuminated by the soft gleam of ambient lights, enhancing the festive mood. The primary focus is on two hands, appearing to belong to an adult of an indeterminate gender and race, engaged in the deliberate act of adorning the Christmas tree. The... | [
""
] | <embedding> | ||
MMEB-V2-R1 | llavahound_caption_retrieval | Narrate the video's unfolding events in a descriptive manner. | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1018553227/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1018553227/c01_0003.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1018553227/c01_0006.jpeg",
"MMEB-train/llavahound/train_video_and_inst... | The video begins with an ambient setting, trees densifying the frame, their languid branches enmeshing the atmosphere. The radiance of the sunrise brilliantly cuts through the muted dew, marking the dawn of a new day. The light begins to fill every nook and cranny, illuminating the richness and diverse textures of the ... | [
""
] | <embedding> | ||
MMEB-V2-R1 | llavahound_caption_retrieval | Offer a detailed interpretation of the video's message and imagery. | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/9139997/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/9139997/c01_0010.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/9139997/c01_0020.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/t... | The video content presents a person engaged in a workout session. Throughout the video, the individual transitions through various poses and exercises. The identifiable attire consists of a light-colored sleeveless top, dark-colored knee-length leggings, and white athletic shoes with a black logo. The background is uni... | [
""
] | <embedding> | ||
MMEB-V2-R1 | llavahound_caption_retrieval | Explain the video's components, including its characters, setting, and plot. | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/v_PWriIL3pTWY-Scene-003/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/v_PWriIL3pTWY-Scene-003/c01_0003.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/v_PWriIL3pTWY-Scene-003/c01_0006.jpeg",
"MME... | The video opens with the individual in a pink costume walking apprehensively around a corner. The setting is an outdoor park with greenery and paved areas. The second individual, wearing only red swim shorts, unexpectedly reveals themselves and a staged pursuit ensues. Both participants demonstrate exaggerated movement... | [
""
] | <embedding> | ||
MMEB-V2-R1 | llavahound_caption_retrieval | Illustrate the video's narrative journey, scene by scene, with attention to detail. | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/A6PZJ4QLQ5U/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/A6PZJ4QLQ5U/c01_0004.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/A6PZJ4QLQ5U/c01_0007.jpeg",
"MMEB-train/llavahound/train_video_and_i... | The video begins with a close-up of a single fingernail painted with a teal green nail polish. A small brush is used to paint a black vertical line down the center of the nail, followed by a horizontal line creating a cross shape. Additional lines are then added to form a simple snowflake pattern. The design is complet... | [
""
] | <embedding> | ||
MMEB-V2-R1 | llavahound_caption_retrieval | Elaborate on the visual and narrative elements of the video in detail. | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/x2M0CZ5Lbks/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/x2M0CZ5Lbks/c01_0005.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/x2M0CZ5Lbks/c01_0009.jpeg",
"MMEB-train/llavahound/train_video_and_i... | The video opens with a sequence of frames showing the ingredients being added to a grinder – chopped bananas, sugar (specific amount is not visible), and chilled milk. The subsequent frames feature chocolate pieces and chocolate syrup being added to the mixture. The shake is then poured into a clear glass, and the fini... | [
""
] | <embedding> | ||
MMEB-V2-R1 | llavahound_caption_retrieval | Analyze the video, focusing on its cinematography and narrative structure. | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/v_eTNHhQxT8/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/v_eTNHhQxT8/c01_0002.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/v_eTNHhQxT8/c01_0003.jpeg",
"MMEB-train/llavahound/train_video_and_i... | The city bus is predominantly blue with yellow and white stripes and accents. The signage on the bus indicates it is the Bx1 service headed towards 165 St. Notable features include the destination sign, reflective mirrors, and patterned tire rims. The exterior of the bus also displays the wheelchair accessibility symbo... | [
""
] | <embedding> | ||
MMEB-V2-R1 | llavahound_video_retrieval | Find a video that contains the following visual content: The video depicts an outdoor environment, possibly a part of an athletic field, as inferred by the presence of tracks and the Olympic rings symbol which indicates a connection to sports. There are multiple black numbers on circular backdrops of different sizes an... | [
""
] | Understand the content of the provided video. | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/v_1Ao43T5yRjI-Scene-004/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/v_1Ao43T5yRjI-Scene-004/c01_0003.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/v_1Ao43T5yRjI-Scene-004/c01_0005.jpeg",
"MME... | <embedding> | ||
MMEB-V2-R1 | llavahound_video_retrieval | Find a video that contains the following visual content: The video provides a close look at the traditional sailing ship, revealing its distinctive features such as the color scheme and the rigging. The hull is painted black, with a strip of green just above the waterline, and the ship has a wooden deck. The sails are ... | [
""
] | Understand the content of the provided video. | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/vulE0iA_98A/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/vulE0iA_98A/c01_0004.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/vulE0iA_98A/c01_0008.jpeg",
"MMEB-train/llavahound/train_video_and_i... | <embedding> | ||
MMEB-V2-R1 | llavahound_video_retrieval | Find a video that contains the following visual content: The video shows a waterfall cascading energetically over a natural rock formation into a body of water below. The water appears frothy and white as it tumbles down, and there is a noticeable mist from the impact. The surroundings are lush with green vegetation, a... | [
""
] | Understand the content of the provided video. | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1017115726/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1017115726/c01_0003.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1017115726/c01_0005.jpeg",
"MMEB-train/llavahound/train_video_and_inst... | <embedding> | ||
MMEB-V2-R1 | llavahound_video_retrieval | Find a video that contains the following visual content: The video begins with a satellite image showing a wider area including Chicago, Indianapolis, and Cincinnati with a measurement marker indicating 77.8 km. The subsequent frames zoom in closer to different areas, showing markers with decreasing distances: 8.91 km,... | [
""
] | Understand the content of the provided video. | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/h-F74IVKBU4/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/h-F74IVKBU4/c01_0006.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/h-F74IVKBU4/c01_0011.jpeg",
"MMEB-train/llavahound/train_video_and_i... | <embedding> | ||
MMEB-V2-R1 | llavahound_video_retrieval | Find a video that contains the following visual content: The series starts with an overview of an indoor soccer pitch, with a player preparing to take a penalty kick. The image is tinted sepia, indicating older footage, likely from a past decade. The player is in a classic soccer kit: white shorts, black socks, and whi... | [
""
] | Understand the content of the provided video. | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/v_vwpaEsh0-1U-Scene-005/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/v_vwpaEsh0-1U-Scene-005/c01_0007.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/v_vwpaEsh0-1U-Scene-005/c01_0014.jpeg",
"MME... | <embedding> | ||
MMEB-V2-R1 | llavahound_video_retrieval | Find a video that contains the following visual content: The video content showcases an animated white rabbit with black eyes and pink ears, appearing to be gender-neutral and of an indeterminate age, stationed to the left of the frame. The rabbit focuses on a detailed anatomical model of the human ear, distinctively o... | [
""
] | Understand the content of the provided video. | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/14545663/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/14545663/c01_0003.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/14545663/c01_0006.jpeg",
"MMEB-train/llavahound/train_video_and_instructio... | <embedding> | ||
MMEB-V2-R1 | llavahound_video_retrieval | Find a video that contains the following visual content: The frames reveal hands skillfully molding and shaping wet clay on a potter's wheel, crafting it into different forms of pottery, such as vases and pots. The potter's hands are covered in the slick, brown clay as they work. The pottery pieces exhibit a range of i... | [
""
] | Understand the content of the provided video. | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/rVEnqrTI1J0/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/rVEnqrTI1J0/c01_0005.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/rVEnqrTI1J0/c01_0009.jpeg",
"MMEB-train/llavahound/train_video_and_i... | <embedding> | ||
MMEB-V2-R1 | llavahound_video_retrieval | Find a video that contains the following visual content: The video contains a series of images depicting a nighttime scene in a snow-covered parking lot. There are a few visible parked cars with their shapes obscured by the dim lighting and falling snow. We see two individuals, one with long hair wearing a golden-color... | [
""
] | Understand the content of the provided video. | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/v_BYLxSOPFOuc-Scene-007/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/v_BYLxSOPFOuc-Scene-007/c01_0004.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/v_BYLxSOPFOuc-Scene-007/c01_0007.jpeg",
"MME... | <embedding> | ||
MMEB-V2-R1 | llavahound_video_retrieval | Find a video that contains the following visual content: The footage shows a person inside a room, dressed in a sports top and shorts, executing a series of exercises. The video kicks off with a still standing pose, transitions to a seated position, and eventually progresses into what is known as 'V Ups,' a core exerci... | [
""
] | Understand the content of the provided video. | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/0N-DJEHet70/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/0N-DJEHet70/c01_0004.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/0N-DJEHet70/c01_0007.jpeg",
"MMEB-train/llavahound/train_video_and_i... | <embedding> | ||
MMEB-V2-R1 | llavahound_video_retrieval | Find a video that contains the following visual content: The video displays a series of still images that depict rolling hills with different color tones ranging from deep green to lighter hues, indicating varying vegetation, with some parts bathed in sunlight creating golden patches. The sky appears mostly cloudy, cas... | [
""
] | Understand the content of the provided video. | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/10057496/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/10057496/c01_0019.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/10057496/c01_0037.jpeg",
"MMEB-train/llavahound/train_video_and_instructio... | <embedding> | ||
MMEB-V2-R1 | llavahound_video_retrieval | Find a video that contains the following visual content: The video features a busy street scene at dusk, with the sky transitioning from blue to darker hues. The street appears wet, possibly after a rainfall, with reflections of streetlights and headlights creating a shimmering effect on the pavement. There are multipl... | [
""
] | Understand the content of the provided video. | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1011113525/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1011113525/c01_0004.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1011113525/c01_0007.jpeg",
"MMEB-train/llavahound/train_video_and_inst... | <embedding> | ||
MMEB-V2-R1 | llavahound_video_retrieval | Find a video that contains the following visual content: The video shows an individual standing in the middle of a pink-walled room with a white ceiling. The room is adorned with red LED lights near the ceiling periphery. They wear an oversized white T-shirt with colorful print design and shorts. The white door behind ... | [
""
] | Understand the content of the provided video. | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/iDJ4DFDioJY/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/iDJ4DFDioJY/c01_0002.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/iDJ4DFDioJY/c01_0004.jpeg",
"MMEB-train/llavahound/train_video_and_i... | <embedding> | ||
MMEB-V2-R1 | llavahound_video_retrieval | Find a video that contains the following visual content: The subject of the video is a young person with short hair. They are wearing a blue t-shirt with a graphic print. The background is minimal and indistinct, mainly a featureless light-colored wall and ceiling. The lighting appears to be artificial, possibly from a... | [
""
] | Understand the content of the provided video. | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/Zed5CxilvJk/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/Zed5CxilvJk/c01_0004.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/Zed5CxilvJk/c01_0008.jpeg",
"MMEB-train/llavahound/train_video_and_i... | <embedding> | ||
MMEB-V2-R1 | llavahound_video_retrieval | Find a video that contains the following visual content: The video presents a series of images where we see an adult male from the shoulders up, appearing to be in a swimming pool. There aren't any drastic changes in the environment or the individual's location throughout the series. In various images, the man's facial... | [
""
] | Understand the content of the provided video. | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/v_tIbSsad0z9U-Scene-002/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/v_tIbSsad0z9U-Scene-002/c01_0003.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/v_tIbSsad0z9U-Scene-002/c01_0005.jpeg",
"MME... | <embedding> | ||
MMEB-V2-R1 | llavahound_video_retrieval | Find a video that contains the following visual content: The footage shows a humpback whale emerging out of the water in a breaching motion. The whale's dark gray body contrasts with the bright blue of the ocean and the green of the trees on the shoreline behind. As the whale breaches, water cascades off its body, crea... | [
""
] | Understand the content of the provided video. | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/WxuMDZPHZSY/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/WxuMDZPHZSY/c01_0003.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/WxuMDZPHZSY/c01_0006.jpeg",
"MMEB-train/llavahound/train_video_and_i... | <embedding> | ||
MMEB-V2-R1 | llavahound_video_retrieval | Find a video that contains the following visual content: The video begins with a shot of the wooden cutting board with water droplets on its surface. The background consists of blurred vegetables. Following this, a hand enters the frame, placing a purple onion on the board and slicing it with a metal knife. The camera ... | [
""
] | Understand the content of the provided video. | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1058125186/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1058125186/c01_0006.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1058125186/c01_0012.jpeg",
"MMEB-train/llavahound/train_video_and_inst... | <embedding> | ||
MMEB-V2-R1 | llavahound_caption_retrieval | Explain the video's components, including its characters, setting, and plot. | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1018427842/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1018427842/c01_0003.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1018427842/c01_0006.jpeg",
"MMEB-train/llavahound/train_video_and_inst... | The footage brings to life a magnificent mountain range, possibly the Rockies or the Alps, though no specific identifiers are present. The impeccably captured video reveals intricate details of the snow-capped peaks, rugged terrain, intricate compositions of rock formations, and the serene yet tumultuous relationship b... | [
""
] | <embedding> | ||
MMEB-V2-R1 | llavahound_caption_retrieval | Elaborate on all aspects of the video you are viewing. | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/9x4WRHraWX0/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/9x4WRHraWX0/c01_0004.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/9x4WRHraWX0/c01_0008.jpeg",
"MMEB-train/llavahound/train_video_and_i... | The video features multiple card images from a football video game, each displaying a player's name, position, national flag (Brazil), and individual skill attributes such as pace (PAC), dribbling (DRI), shooting (SHO), defense (DEF), passing (PAS), and physicality (PHY). Each player card also has an overall rating out... | [
""
] | <embedding> | ||
MMEB-V2-R1 | llavahound_caption_retrieval | Elaborate on the visual and narrative elements of the video in detail. | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1014429143/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1014429143/c01_0003.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1014429143/c01_0006.jpeg",
"MMEB-train/llavahound/train_video_and_inst... | The video shows a close-up of an adult human hand, with no distinguishing gender features, manipulating multiple United States one-dollar bills. The hand appears to be Caucasian as evidenced by the lighter skin tone. The identifiable one-dollar bills are notably printed with the number "1" and bear the Great Seal of th... | [
""
] | <embedding> | ||
MMEB-V2-R1 | llavahound_caption_retrieval | Provide an exhaustive description of the video content. | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/OWNv3uwwl4A/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/OWNv3uwwl4A/c01_0005.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/OWNv3uwwl4A/c01_0010.jpeg",
"MMEB-train/llavahound/train_video_and_i... | The video starts with a hand holding materials needed for making pearl earrings against a purple background. The materials include golden chains, rings, hooks, and white pearls. The hand proceeds to show golden chains of varying lengths and a pair of hand tools, implying the process of crafting. Throughout the video, t... | [
""
] | <embedding> | ||
MMEB-V2-R1 | llavahound_caption_retrieval | Explain the video's components, including its characters, setting, and plot. | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1033637198/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1033637198/c01_0009.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1033637198/c01_0017.jpeg",
"MMEB-train/llavahound/train_video_and_inst... | The video features a sequence of images showcasing concentric circles made up of small white dots, which give the appearance of being illuminated. The background is a deep blue, which enhances the contrast of the white dots. The pattern of the dots remains uniform throughout the video, creating a hypnotic, swirling eff... | [
""
] | <embedding> | ||
MMEB-V2-R1 | llavahound_caption_retrieval | Provide a comprehensive analysis of the video's content and themes. | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/uBf0H0HkjCY/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/uBf0H0HkjCY/c01_0004.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/uBf0H0HkjCY/c01_0007.jpeg",
"MMEB-train/llavahound/train_video_and_i... | The first frame of the video features a cricket pitch with a batsman at the crease facing a bowler in mid-action, about to deliver the ball. An umpire is also present in the frame. There are advertisements such as "TATA PUNCH" on the screen's background, consistent with branding commonly found in sports stadiums. The s... | [
""
] | <embedding> | ||
MMEB-V2-R1 | llavahound_caption_retrieval | Elaborate on all aspects of the video you are viewing. | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/IafENDrONBM/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/IafENDrONBM/c01_0005.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/IafENDrONBM/c01_0009.jpeg",
"MMEB-train/llavahound/train_video_and_i... | The frames feature a display of various toy products, with a prominent focus on the "LOL OMG" dolls in boxed packaging. These boxes have a distinct black and white stripe pattern with bright pink accents and showcase illustrations of the dolls dressed in stylish outfits. The price of $9.99 is visible. The shelves are m... | [
""
] | <embedding> | ||
MMEB-V2-R1 | llavahound_caption_retrieval | Offer a detailed interpretation of the video's message and imagery. | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/v_ow9bWn5gOvg-Scene-013/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/v_ow9bWn5gOvg-Scene-013/c01_0003.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/v_ow9bWn5gOvg-Scene-013/c01_0005.jpeg",
"MME... | The person, appearing to be a chef given their white coat, is working at a wooden cutting board located in a well-lit kitchen. The environment suggests a home-style setup with a green bowl, various kitchenware, and a cooker visible in the frame. The chef's actions involve handling and possibly chopping ingredients, tho... | [
""
] | <embedding> | ||
MMEB-V2-R1 | llavahound_qa | Answer a question based on the content of a video.
What is the purpose of the paper plates with money and the different labels on them? | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/Lftiu81VDM0/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/Lftiu81VDM0/c01_0005.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/Lftiu81VDM0/c01_0010.jpeg",
"MMEB-train/llavahound/train_video_and_i... | The purpose of the paper plates with money and different labels on them is to collect donations representative of each religion's concern for the homeless. | [
""
] | <embedding> | ||
MMEB-V2-R1 | llavahound_qa | Answer a question based on the content of a video.
What is the setting of the video? | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1035577184/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1035577184/c01_0003.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1035577184/c01_0005.jpeg",
"MMEB-train/llavahound/train_video_and_inst... | The setting of the video appears to be a residential area with suburban homes or properties visible in the blurred backdrop. | [
""
] | <embedding> | ||
MMEB-V2-R1 | llavahound_qa | Answer a question based on the content of a video.
How is the Betta fish's ascent characterized in the video? | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1026980378/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1026980378/c01_0003.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1026980378/c01_0006.jpeg",
"MMEB-train/llavahound/train_video_and_inst... | The Betta fish's ascent is characterized in the video by fluidity and calmness. | [
""
] | <embedding> | ||
MMEB-V2-R1 | llavahound_qa | Answer a question based on the content of a video.
What type of lawn mower is being used in the video? | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/v_sanasW2azw4-Scene-002/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/v_sanasW2azw4-Scene-002/c01_0012.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/v_sanasW2azw4-Scene-002/c01_0023.jpeg",
"MME... | The lawn mower appears to be a standard gas-powered model with wheels, a handle, and a base housing the engine and cutting blade. | [
""
] | <embedding> | ||
MMEB-V2-R1 | llavahound_qa | Answer a question based on the content of a video.
What is the individual wearing a green shirt and dark sunglasses doing in the subsequent frames? | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/55dAKLE8Pw8/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/55dAKLE8Pw8/c01_0002.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/55dAKLE8Pw8/c01_0004.jpeg",
"MMEB-train/llavahound/train_video_and_i... | The individual in the subsequent frames seems to be speaking or gesturing to someone out of frame. | [
""
] | <embedding> | ||
MMEB-V2-R1 | llavahound_qa | Answer a question based on the content of a video.
What is a polygraph machine? | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/761941/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/761941/c01_0003.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/761941/c01_0006.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/trai... | A polygraph machine is a tool often employed to ascertain an individual's truthfulness by revealing visible fluctuations in physiological attributes. | [
""
] | <embedding> | ||
MMEB-V2-R1 | llavahound_qa | Answer a question based on the content of a video.
What is the main focus of the series of images? | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1025948129/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1025948129/c01_0008.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1025948129/c01_0016.jpeg",
"MMEB-train/llavahound/train_video_and_inst... | The main focus of the series of images is a plant with multiple barren branches extending upward from a central base. | [
""
] | <embedding> | ||
MMEB-V2-R1 | llavahound_qa | Answer a question based on the content of a video.
What is the focal point of the kitchen scene? | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/v_TDWbRqIEcts-Scene-006/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/v_TDWbRqIEcts-Scene-006/c01_0003.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/v_TDWbRqIEcts-Scene-006/c01_0005.jpeg",
"MME... | The focal point of the kitchen scene is a ceramic mug and a glass of milk positioned side by side on a countertop. | [
""
] | <embedding> | ||
MMEB-V2-R1 | llavahound_video_retrieval | Find a video that contains the following visual content: The person in focus is performing ice skating maneuvers on a large, open-air ice rink. They are wearing black clothing and white ice skates. They exhibit intermediate to advanced skating skills, executing turns, spins, and a split jump. The environment around the... | [
""
] | Understand the content of the provided video. | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/REcCgk_ZSck/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/REcCgk_ZSck/c01_0002.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/REcCgk_ZSck/c01_0003.jpeg",
"MMEB-train/llavahound/train_video_and_i... | <embedding> | ||
MMEB-V2-R1 | llavahound_video_retrieval | Find a video that contains the following visual content: This video sequence consists of similar clips that show a light beige ball of dough resting on a circular wooden board covered in a layer of white flour. To the left of the board, a wooden rolling pin lies idle, and behind the working area, there is a clear glass... | [
""
] | Understand the content of the provided video. | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1042914457/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1042914457/c01_0008.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1042914457/c01_0016.jpeg",
"MMEB-train/llavahound/train_video_and_inst... | <embedding> | ||
MMEB-V2-R1 | llavahound_video_retrieval | Find a video that contains the following visual content: The video starts with a view of the water reflecting the cloudy sky as the camera pans along the walking path lined by rocks on one side and water on the other. We see the reflection of trees along the water's edge and different angles of the surrounding environm... | [
""
] | Understand the content of the provided video. | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/Oylx7NmzTDo/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/Oylx7NmzTDo/c01_0003.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/Oylx7NmzTDo/c01_0006.jpeg",
"MMEB-train/llavahound/train_video_and_i... | <embedding> | ||
MMEB-V2-R1 | llavahound_video_retrieval | Find a video that contains the following visual content: The video exclusively features a stationary shot of an escalator in operation without any human activity or inanimate objects. The escalator consists of metal steps with black ribbed tractions and yellow demarcation lines on either side of each step for safety. T... | [
""
] | Understand the content of the provided video. | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1027812956/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1027812956/c01_0003.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1027812956/c01_0006.jpeg",
"MMEB-train/llavahound/train_video_and_inst... | <embedding> | ||
MMEB-V2-R1 | llavahound_video_retrieval | Find a video that contains the following visual content: In the video, a clear glass mug is set on the kitchen counter with a white-tiled wall and a speckled countertop visible in the background. An individual begins by pouring a white liquid, presumably milk, into the mug. They then add a green, matcha-based liquid to... | [
""
] | Understand the content of the provided video. | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/tAhWDeXCrvE/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/tAhWDeXCrvE/c01_0005.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/tAhWDeXCrvE/c01_0009.jpeg",
"MMEB-train/llavahound/train_video_and_i... | <embedding> | ||
MMEB-V2-R1 | llavahound_video_retrieval | Find a video that contains the following visual content: The video shows a single gas stove burner igniting and extinguishing, captured against a black background. The flame displays hues of blue and a hint of red near the base. The circular burner consists of several small ports from which the gas flame emerges. The s... | [
""
] | Understand the content of the provided video. | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/33066460/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/33066460/c01_0004.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/33066460/c01_0007.jpeg",
"MMEB-train/llavahound/train_video_and_instructio... | <embedding> | ||
MMEB-V2-R1 | llavahound_video_retrieval | Find a video that contains the following visual content: The video includes multiple frames where the demonstrator, wearing a white t-shirt, black pants, and red boxing shoes, showcases the technique. The boxing bag is black and stands in a room with visible heavy bags in the background. The video uses visual annotatio... | [
""
] | Understand the content of the provided video. | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/MAcpjnO2nmE/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/MAcpjnO2nmE/c01_0005.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/MAcpjnO2nmE/c01_0009.jpeg",
"MMEB-train/llavahound/train_video_and_i... | <embedding> | ||
MMEB-V2-R1 | llavahound_video_retrieval | Find a video that contains the following visual content: In detail, a pair of Caucasian hands, presumably belonging to an adult of unidentifiable gender, engages in the art of sewing. The individual is operating a white sewing machine, a model unspecified but equipped with a compact yet efficient work area illuminated ... | [
""
] | Understand the content of the provided video. | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1027854194/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1027854194/c01_0003.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1027854194/c01_0006.jpeg",
"MMEB-train/llavahound/train_video_and_inst... | <embedding> | ||
MMEB-V2-R1 | llavahound_caption_retrieval | Describe every scene and its significance in the video. | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/Ywt1kz4IxcQ/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/Ywt1kz4IxcQ/c01_0005.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/Ywt1kz4IxcQ/c01_0009.jpeg",
"MMEB-train/llavahound/train_video_and_i... | The individual is dressed in a red t-shirt and red trousers. The room has a modern decor with decorative lights around the perimeter, a television mounted on the wall, and a white fluffy wall texture to the side. There is also a clay pot on the floor and a reflective surface, likely a mirror, on the back wall. The vide... | [
""
] | <embedding> | ||
MMEB-V2-R1 | llavahound_caption_retrieval | Dissect the video's content, explaining each element thoroughly. | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/v_R547-iOgb6g-Scene-007/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/v_R547-iOgb6g-Scene-007/c01_0006.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/v_R547-iOgb6g-Scene-007/c01_0011.jpeg",
"MME... | The video consists of a repeated series of images, where the first image is of a plain white background followed by images that showcase two distinct types of mop sleeves with the "UNGER" logo and "Quality Tools for Smart Cleaning" tagline present at the top. One mop sleeve is fully red, while the other has red materia... | [
""
] | <embedding> | ||
MMEB-V2-R1 | llavahound_caption_retrieval | Explain the video's components, including its characters, setting, and plot. | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/7oFdEWOlF4w/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/7oFdEWOlF4w/c01_0003.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/7oFdEWOlF4w/c01_0006.jpeg",
"MMEB-train/llavahound/train_video_and_i... | The video shows close-up shots of a person's hand holding a yellow and white condensed milk can with a logo featuring a spoon and swirl design. They are drizzling a thick stream of off-white condensed milk onto a golden-brown, rectangular piece of fried bread in a dark-colored frying pan. The bread appears crispy and h... | [
""
] | <embedding> | ||
MMEB-V2-R1 | llavahound_caption_retrieval | Explain the video's components, including its characters, setting, and plot. | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/VwFQxCGHAUY/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/VwFQxCGHAUY/c01_0002.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/VwFQxCGHAUY/c01_0004.jpeg",
"MMEB-train/llavahound/train_video_and_i... | In the video, the person, attired in a black shirt with a bright green circular emblem in the centre and black pants, executes a skateboarding maneuver. The skateboard has a white underbelly with hints of a green design similar to the shirt's emblem, and the shoes are turqoise with white soles. The setting is a public,... | [
""
] | <embedding> | ||
MMEB-V2-R1 | llavahound_caption_retrieval | Convey the narrative and visual elements of the video in detail. | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/5936804/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/5936804/c01_0003.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/5936804/c01_0005.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/t... | The video reveals a scene where chefs, dressed in professional white attire, are engaged in the careful presentation of desserts. Their attire comprises neat chef whites, indicating a professional environment, perhaps a restaurant or a patisserie. One chef is seen using a spoon, skillfully adding a finishing touch to a... | [
""
] | <embedding> | ||
MMEB-V2-R1 | llavahound_caption_retrieval | Write an in-depth depiction of the video, covering all its aspects. | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1023999554/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1023999554/c01_0003.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1023999554/c01_0006.jpeg",
"MMEB-train/llavahound/train_video_and_inst... | The individual in the video is wearing a formal grey suit with a blue tie. They are holding a microphone in their right hand and using their left hand to gesture as they speak. The person is standing against a plain, light-colored background. There are no discernible logos or brands visible in the frames provided. The ... | [
""
] | <embedding> | ||
MMEB-V2-R1 | llavahound_caption_retrieval | Elaborate on all aspects of the video you are viewing. | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1022053387/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1022053387/c01_0003.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1022053387/c01_0006.jpeg",
"MMEB-train/llavahound/train_video_and_inst... | The video is readily recognizable as a sales advertisement showcasing a vibrant "SALE" signage positioned centrally. It undergoes a captivating transformation, alternating between the colors green, blue, and pink, creating an alluring visual effect designed to capture and sustain viewers’ attention. The background is i... | [
""
] | <embedding> | ||
MMEB-V2-R1 | llavahound_caption_retrieval | Elaborate on all aspects of the video you are viewing. | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/37v_cVszXtM/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/37v_cVszXtM/c01_0002.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/37v_cVszXtM/c01_0004.jpeg",
"MMEB-train/llavahound/train_video_and_i... | The video opens with a sketchbook on a wooden desk. Surrounding the sketchbook are a black pen, plants in pots, notebooks, and ornaments. The sketchbook contains detailed drawings featuring figures with distinct attire and dripping elements, resembling ink or paint. The style is monochrome with heavy shading. Text acco... | [
""
] | <embedding> | ||
MMEB-V2-R1 | llavahound_qa | Answer a question based on the content of a video.
What are the concentric circles known as, that are marked on the curling sheet? | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/v_JBnL0T2XzGs-Scene-001/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/v_JBnL0T2XzGs-Scene-001/c01_0005.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/v_JBnL0T2XzGs-Scene-001/c01_0009.jpeg",
"MME... | Each curling sheet is marked with concentric circles known as the house. | [
""
] | <embedding> | ||
MMEB-V2-R1 | llavahound_qa | Answer a question based on the content of a video.
What do the boats in the video appear to be made of? | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1062388561/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1062388561/c01_0005.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1062388561/c01_0009.jpeg",
"MMEB-train/llavahound/train_video_and_inst... | The boats in the video appear to be made of wood. | [
""
] | <embedding> | ||
MMEB-V2-R1 | llavahound_qa | Answer a question based on the content of a video.
What type of clothes is the person wearing in the video? | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/v_fkgojk18K0E-Scene-001/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/v_fkgojk18K0E-Scene-001/c01_0013.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/v_fkgojk18K0E-Scene-001/c01_0025.jpeg",
"MME... | The person in the video is wearing belly dancing attire, which consists of a black top, black pants, and a white hip scarf with decorative elements. | [
""
] | <embedding> | ||
MMEB-V2-R1 | llavahound_qa | Answer a question based on the content of a video.
What is the main focus of the video? | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/Dr5bsRy8p8o/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/Dr5bsRy8p8o/c01_0005.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/Dr5bsRy8p8o/c01_0009.jpeg",
"MMEB-train/llavahound/train_video_and_i... | The main focus of the video is a towering white high-rise building with multiple rows of windows and balconies. | [
""
] | <embedding> | ||
MMEB-V2-R1 | llavahound_qa | Answer a question based on the content of a video.
What is the recurring graphical element in the video? | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/gyCz8uL02GA/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/gyCz8uL02GA/c01_0006.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/gyCz8uL02GA/c01_0011.jpeg",
"MMEB-train/llavahound/train_video_and_i... | A recurring graphical element in the video is a cartoon rabbit holding an Easter basket, positioned in the bottom left corner. | [
""
] | <embedding> | ||
MMEB-V2-R1 | llavahound_qa | Answer a question based on the content of a video.
How is nature showcased in this video? | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1008119554/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1008119554/c01_0003.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/1008119554/c01_0006.jpeg",
"MMEB-train/llavahound/train_video_and_inst... | Nature is shown in its raw and unbridled form with the tempest of tornado, which grows menacingly and expands with dark tendrils of clouds weaving and coiling in a devastating dance. The storm's rapid birth and growth is as terrifying as it is mesmerizing. | [
""
] | <embedding> | ||
MMEB-V2-R1 | llavahound_qa | Answer a question based on the content of a video.
What does the aerial view reveal about the solar farm? | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/19553689/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/19553689/c01_0012.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/19553689/c01_0023.jpeg",
"MMEB-train/llavahound/train_video_and_instructio... | The aerial view reveals an expansive solar farm consisting of numerous rows of blue photovoltaic panels that are neatly arranged in parallel lines. | [
""
] | <embedding> | ||
MMEB-V2-R1 | llavahound_qa | Answer a question based on the content of a video.
What is the atmosphere of the kitchen like in the video? | [
"MMEB-train/llavahound/train_video_and_instruction/train_300k/v_-zp0Sban5Wk-Scene-013/c01_0001.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/v_-zp0Sban5Wk-Scene-013/c01_0002.jpeg",
"MMEB-train/llavahound/train_video_and_instruction/train_300k/v_-zp0Sban5Wk-Scene-013/c01_0004.jpeg",
"MME... | The ambient lighting is warm and consistent, which gives the kitchen a homey and inviting atmosphere. | [
""
] | <embedding> | ||
MMEB-V2-R1 | openbmb/VisRAG-Ret-Train-In-domain-data | This query is related to retrieving an infographic that visually presents statistical or factual information using charts, icons, and structured layouts. The retrieved image should contain the necessary visual elements to provide the best context for answering the query: How many states report more COVID-19 cases than ... | [
""
] | An infographic with structured data, charts, and annotations. | MMEB-train/visrag/images/81631.jpg | [
""
] | <embedding> | |
MMEB-V2-R1 | openbmb/VisRAG-Ret-Train-In-domain-data | This query is related to retrieving a relevant figure from an ArXiv research paper. The retrieved figure should contain scientific plots, mathematical visualizations, or experimental results that best address the query: Based on the bottom figure, which statement is true about the transformation? | [
""
] | A figure from a research paper, including plots or experimental results. | MMEB-train/visrag/images/29948.jpg | [
""
] | <embedding> | |
MMEB-V2-R1 | openbmb/VisRAG-Ret-Train-In-domain-data | This query relates to retrieving a relevant plot or chart that visually represents numerical data. The retrieved figure should contain the necessary information to analyze trends, compare values, or extract key insights related to the query: What does the 2nd bar from the top in Haiti represents ? | [
""
] | A numerical data visualization, such as bar charts or line graphs. | MMEB-train/visrag/images/25913.png | [
""
] | <embedding> | |
MMEB-V2-R1 | openbmb/VisRAG-Ret-Train-In-domain-data | This query is related to retrieving a relevant figure from an ArXiv research paper. The retrieved figure should contain scientific plots, mathematical visualizations, or experimental results that best address the query: Based on figures (b) and (c), what can be inferred about the distribution of betweenness centrality ... | [
""
] | A figure from a research paper, including plots or experimental results. | MMEB-train/visrag/images/46830.jpg | [
""
] | <embedding> | |
MMEB-V2-R1 | openbmb/VisRAG-Ret-Train-In-domain-data | This query relates to retrieving a relevant plot or chart that visually represents numerical data. The retrieved figure should contain the necessary information to analyze trends, compare values, or extract key insights related to the query: What is the median long-term female unemployment ? | [
""
] | A numerical data visualization, such as bar charts or line graphs. | MMEB-train/visrag/images/68745.png | [
""
] | <embedding> | |
MMEB-V2-R1 | openbmb/VisRAG-Ret-Train-In-domain-data | This query relates to retrieving a relevant plot or chart that visually represents numerical data. The retrieved figure should contain the necessary information to analyze trends, compare values, or extract key insights related to the query: What is the difference between the highest and the second highest earnings fro... | [
""
] | A numerical data visualization, such as bar charts or line graphs. | MMEB-train/visrag/images/91880.png | [
""
] | <embedding> | |
MMEB-V2-R1 | openbmb/VisRAG-Ret-Train-In-domain-data | This query relates to retrieving a relevant plot or chart that visually represents numerical data. The retrieved figure should contain the necessary information to analyze trends, compare values, or extract key insights related to the query: How many bars are there on the 3rd tick from the left ? | [
""
] | A numerical data visualization, such as bar charts or line graphs. | MMEB-train/visrag/images/110462.png | [
""
] | <embedding> | |
MMEB-V2-R1 | openbmb/VisRAG-Ret-Train-In-domain-data | This query is related to retrieving a relevant page from a multi-page document, such as reports, invoices, or research papers. The retrieved document should contain text, tables, or structured information necessary to answer the query: What is the profit-sharing ratio of Sanjay J. Desai? | [
""
] | A page from a multi-page document with text or tables. | MMEB-train/visrag/images/83490.jpg | [
""
] | <embedding> | |
MMEB-V2-R1 | vidore/colpali_train_set | Represent the given image with the following question: What degree is being pursued with this dissertation? | MMEB-train/vidore/images/100772.jpg | [
""
] | ['Doctor of Philosophy in Electrical Engineering'] | [
""
] | <embedding> | |
MMEB-V2-R1 | vidore/colpali_train_set | Represent the given image with the following question: What is the full form of PNGV?
Offer a very short reply. | MMEB-train/vidore/images/108412.jpg | [
""
] | Partnership for New Generation of Vehicles. | [
""
] | <embedding> | |
MMEB-V2-R1 | vidore/colpali_train_set | Represent the given image with the following question: For what fiscal year is this report prepared? | MMEB-train/vidore/images/8394.jpg | [
""
] | ['Fiscal year ended September 30, 2021'] | [
""
] | <embedding> | |
MMEB-V2-R1 | vidore/colpali_train_set | Represent the given image with the following question: What resources are provided for comprehensive and integrated chronic disease prevention? | MMEB-train/vidore/images/83831.jpg | [
""
] | ['Comprehensive and Integrated Chronic Disease Prevention: Action Planning Handbook for States and Communities by National Association of Chronic Disease Directors. 2005.', 'Recommendations for Integration of Chronic Disease Programs: Are Your Programs Linked? Prev Chronic Dis [serial online] 2007.'] | [
""
] | <embedding> | |
MMEB-V2-R1 | vidore/colpali_train_set | Represent the given image with the following question: What tasks can Gradient Boosting perform? | MMEB-train/vidore/images/221.jpg | [
""
] | ['regression', 'classification', 'ranking'] | [
""
] | <embedding> | |
MMEB-V2-R1 | vidore/colpali_train_set | Represent the given image with the following question: What was the respective research and development expense in 2019, 2018 and 2017? | MMEB-train/vidore/images/101394.jpg | [
""
] | ['16,545', '17,953', '16,190'] | [
""
] | <embedding> | |
MMEB-V2-R1 | vidore/colpali_train_set | Represent the given image with the following question: What subjects are included in the data and for what age ranges? | MMEB-train/vidore/images/109838.jpg | [
""
] | ['The subjects included are Educ, Abil, Math, Sci, Read, and Avg. The age ranges cover 7-80 years old.'] | [
""
] | <embedding> | |
MMEB-V2-R1 | vidore/colpali_train_set | Represent the given image with the following question: In hypothetical combinations what is the parts per million value of sodium chloride ?
Provide a succinct answer. | MMEB-train/vidore/images/81158.jpg | [
""
] | 30. | [
""
] | <embedding> | |
MMEB-V2-R1 | vidore/colpali_train_set | Represent the given image with the following question: Which organization produced this document? | MMEB-train/vidore/images/116646.jpg | [
""
] | ['Residential Information Systems Project (RISP)', 'Research and Training Center on Community Living', 'Institute on Community Integration/UCEDD', 'UNIVERSITY OF MINNESOTA'] | [
""
] | <embedding> | |
MMEB-V2-R1 | vidore/colpali_train_set | Represent the given image with the following question: What is the ideal location for a site in the hub-and-spoke model? | MMEB-train/vidore/images/58369.jpg | [
""
] | ['Near a major highway or interstate for access to ground transportation', 'Co-located with an operational airport', 'Near the impacted area (e.g., within an hour) without impeding response efforts'] | [
""
] | <embedding> | |
MMEB-V2-R1 | vidore/colpali_train_set | Represent the given image with the following question: What operation on GPU is represented as having the fastest execution time for \( X_{max} = 2048 \) based on the figure? | MMEB-train/vidore/images/50612.jpg | [
""
] | C. QR+CBE | [
""
] | <embedding> | |
MMEB-V2-R1 | vidore/colpali_train_set | Represent the given image with the following question: According to the research participant, what are the benefits of bringing theory back into research? | MMEB-train/vidore/images/100180.jpg | [
""
] | ['it results in better questions', 'it might have more impact because diffused findings could be attached to a larger spread of understanding', 'it could potentially bring new methods'] | [
""
] | <embedding> | |
MMEB-V2-R1 | vidore/colpali_train_set | Represent the given image with the following question: What percentage of design contracts had missing information according to the data? | MMEB-train/vidore/images/101311.jpg | [
""
] | ['25%'] | [
""
] | <embedding> | |
MMEB-V2-R1 | vidore/colpali_train_set | Represent the given image with the following question: What role did the literature review play in the research? | MMEB-train/vidore/images/45796.jpg | [
""
] | ['critical in identifying risky communicable diseases', 'suggested numerous legal and ethical guidelines a health care provider must follow'] | [
""
] | <embedding> | |
MMEB-V2-R1 | vidore/colpali_train_set | Represent the given image with the following question: What was the change in trade receivables in 2019 from 2018? | MMEB-train/vidore/images/42638.jpg | [
""
] | 5.3 | [
""
] | <embedding> | |
MMEB-V2-R1 | vidore/colpali_train_set | Represent the given image with the following question: What are the factors that determine the economic competitiveness of nuclear power in a country? | MMEB-train/vidore/images/13205.jpg | [
""
] | ['alternatives available', 'overall electricity demand and growth', 'market structure and investment environment'] | [
""
] | <embedding> | |
MMEB-V2-R1 | openbmb/VisRAG-Ret-Train-In-domain-data | This query is related to retrieving an infographic that visually presents statistical or factual information using charts, icons, and structured layouts. The retrieved image should contain the necessary visual elements to provide the best context for answering the query: How many times twitter users tweet every minute ... | [
""
] | An infographic with structured data, charts, and annotations. | MMEB-train/visrag/images/48598.jpg | [
""
] | <embedding> | |
MMEB-V2-R1 | openbmb/VisRAG-Ret-Train-In-domain-data | This query relates to retrieving a relevant plot or chart that visually represents numerical data. The retrieved figure should contain the necessary information to analyze trends, compare values, or extract key insights related to the query: How many bars are there on the 3rd tick from the left ? | [
""
] | A numerical data visualization, such as bar charts or line graphs. | MMEB-train/visrag/images/96204.png | [
""
] | <embedding> | |
MMEB-V2-R1 | openbmb/VisRAG-Ret-Train-In-domain-data | This query is related to retrieving a relevant presentation slide that visually presents structured information. The retrieved slide should contain the necessary text, charts, or graphics to provide the best answer to the query: Which version of that which is used by 47% of the popular sites on the web is used the most... | [
""
] | A presentation slide with text, bullet points, and diagrams. | MMEB-train/visrag/images/25044.jpg | [
""
] | <embedding> | |
MMEB-V2-R1 | openbmb/VisRAG-Ret-Train-In-domain-data | This query relates to retrieving a relevant plot or chart that visually represents numerical data. The retrieved figure should contain the necessary information to analyze trends, compare values, or extract key insights related to the query: How many legend labels are there ? | [
""
] | A numerical data visualization, such as bar charts or line graphs. | MMEB-train/visrag/images/63559.png | [
""
] | <embedding> | |
MMEB-V2-R1 | openbmb/VisRAG-Ret-Train-In-domain-data | This query is related to retrieving an infographic that visually presents statistical or factual information using charts, icons, and structured layouts. The retrieved image should contain the necessary visual elements to provide the best context for answering the query: What are the three colours on German flag- black... | [
""
] | An infographic with structured data, charts, and annotations. | MMEB-train/visrag/images/67110.jpg | [
""
] | <embedding> | |
MMEB-V2-R1 | openbmb/VisRAG-Ret-Train-In-domain-data | This query is related to retrieving a relevant presentation slide that visually presents structured information. The retrieved slide should contain the necessary text, charts, or graphics to provide the best answer to the query: What is the investment size of a seed in millions of euro? | [
""
] | A presentation slide with text, bullet points, and diagrams. | MMEB-train/visrag/images/102526.jpg | [
""
] | <embedding> | |
MMEB-V2-R1 | openbmb/VisRAG-Ret-Train-In-domain-data | This query is related to retrieving an infographic that visually presents statistical or factual information using charts, icons, and structured layouts. The retrieved image should contain the necessary visual elements to provide the best context for answering the query: Which all are the cities in Canada with fast gro... | [
""
] | An infographic with structured data, charts, and annotations. | MMEB-train/visrag/images/107239.jpg | [
""
] | <embedding> | |
MMEB-V2-R1 | openbmb/VisRAG-Ret-Train-In-domain-data | This query is related to retrieving a relevant figure from an ArXiv research paper. The retrieved figure should contain scientific plots, mathematical visualizations, or experimental results that best address the query: What does the parameter \( \omega/\omega_0 \) likely represent in the graph? | [
""
] | A figure from a research paper, including plots or experimental results. | MMEB-train/visrag/images/1173.jpg | [
""
] | <embedding> | |
MMEB-V2-R1 | vidore/colpali_train_set | Represent the given image with the following question: What actions should be taken if the mean CT number or noise deviates from the QMP criteria? | MMEB-train/vidore/images/14293.jpg | [
""
] | ['phantom', 'phantom positioning', 'phantom image used', 'ROI placement', 'protocol used should be double checked', 'air calibrations should be run', 'test should be repeated'] | [
""
] | <embedding> | |
MMEB-V2-R1 | vidore/colpali_train_set | Represent the given image with the following question: What is the status/action of P3(7)?
Make the answer very short. | MMEB-train/vidore/images/45967.jpg | [
""
] | Manuscript in preparation. | [
""
] | <embedding> | |
MMEB-V2-R1 | vidore/colpali_train_set | Represent the given image with the following question: What approach do many students take after receiving a poor test score that makes it difficult to learn from their mistakes? | MMEB-train/vidore/images/52669.jpg | [
""
] | ['avoiding looking at the test in detail', 'trying to forget about the unpleasant experience'] | [
""
] | <embedding> | |
MMEB-V2-R1 | vidore/colpali_train_set | Represent the given image with the following question: What is the date mentioned in the top of the document ?
Offer a terse response. | MMEB-train/vidore/images/42629.jpg | [
""
] | 1-29-69. | [
""
] | <embedding> |
End of preview. Expand in Data Studio
UME-rl-train
Introduction
The dataset is constructed based on the training set of MMEB-V2. The dataset is evenly sampled from the MMEB-V2 dataset, which covers image, video, and visual document modalities, ensuring that the total amount of data across different modalities is approximately balanced, as well as the amount of data within each dataset under the same modality. This results in a final set of 11,136 RL pairs.
Citation
If you find our work useful, please consider citing it.
@article{lan2025ume,
title={UME-R1: Exploring Reasoning-Driven Generative Multimodal Embeddings},
author={Lan, Zhibin and Niu, Liqiang and Meng, Fandong and Zhou, Jie and Su, Jinsong},
journal={arXiv preprint arXiv:2511.00405},
year={2025}
}
- Downloads last month
- 33