Panoramic 360 Stereo Pairs for Depth Estimation
39,553 real-world panoramic stereo images with depth maps and LiDAR disparity.
- Depth estimation
Hugging Face datasets with plain-language descriptions: what's inside and what models to train on them. 4,788 in the catalog so far, with new ones added every hour.
Answered by Gemini 4 · the chat is kept only in this tab
39,553 real-world panoramic stereo images with depth maps and LiDAR disparity.
Generated images of people in various outfits for virtual try-on tasks.
Camera, lidar, radar, and microphone recordings for detecting and tracking miniature UAVs.
A dynamic English-language benchmark for assessing how well multimodal LLMs understand scientific journals and illustrations.
Four LIBERO datasets with Panda robot episodes, merged and converted to the LeRobot format.
A set of under 1,000 complex multimodal questions with images for testing the limits of modern models.
Formatted DocVQA version for one-click evaluation of large multimodal models.
35,000 images from the GQA dataset in FiftyOne format with object and attribute scene graph annotations, but without questions.
Kubric-generated videos with ground-truth 3D point trajectories for training trackers.
Generated pages in Vietnamese and English with block annotations for training document parsing.
3,493 medical ultrasound images of the thyroid with high-quality annotations for segmentation and classification tasks.
Regularly updated corpus of posts, comments, photos, videos, and profiles from VK in Russian and Ukrainian.
Formatted version of the TextVQA benchmark for evaluating multimodal models in lmms-eval.
Set of queries and mixed corpora based on MSCOCO, Google_WIT, VisualNews, and OVEN for cross-modal search evaluation.
A benchmark for evaluating real enterprise document parsing systems across five dimensions.
A large-scale multimodal benchmark for evaluating action understanding in video.
Multimodal dataset from Amazon Science to test VLM capability to find answers in bulky documents and images.
A processed and combined set of MNIST images.
A small image-format climate dataset for extreme compression of ERA5-level archives.
2,554 episodes in AI2-THOR and ProcTHOR simulators for evaluating the memory of embodied agents.
A benchmark of 1,651 PDF pages for evaluating document parsing across various types and languages.
Bird's-eye view frames from the AirSim simulator for depth estimation and segmentation.
Mathematical reasoning benchmark over images: IQTest, FunctionQA, and PaperQA; languages en, zh, fa.
Over 15 thousand English and Chinese prompts with metadata and generated images from GPT Image 2.