MacroSony/inat-2017-subset

iNaturalist 2017 Supercategory Subset This dataset is a sampled subset of the iNaturalist 2017 Challenge dataset, specifically processed for efficient object detection training across major biological supercategories. Dataset Summary Source: iNaturalist 2017 (Competition Version) Task: Object Detection Classes: 9 (Collapsed from thousands of species into biological supercategories) Training Samples: 1,000 images per supercategory (~9,000 total) Validation… See the full description on the dataset page: https://huggingface.co/datasets/MacroSony/inat-2017-subset.