datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
Agricultural-Pest-Disease-Image-Annotation-Bounding-Box-Polygon-Detection
Agricultural Pest & Disease Image Annotation (Bounding Box/Polygon Detection)
For intelligent pest and disease recognition in the agriculture and livestock industry, this dataset provides target region annotations on raw plant/disease images. Annotation results support both bounding boxes and polygon shapes, enabling models to learn object locations and fine-grained contours. It covers common pest/disease targets and supplies strong supervision for detection and segmentation… See the full description on the dataset page: https://huggingface.co/datasets/shangzx/Agricultural-Pest-Disease-Image-Annotation-Bounding-Box-Polygon-Detection.scent-engine-annotations
ScentEngine — Universal Training Annotations
Two-stage Gemma 4 labels (E4B primary, 31B verifier for conf ∈ [0.4, 0.7])
over public game screenshots. Each row maps an image to 6 PWM values
(0–100) in locked channel order: Flora · Aqua · Earth · Pyric · Ozone · Civic.
Files
File
Purpose
public_games_final.jsonl
All accepted labels (quarantine excluded)
universal_train.jsonl
90% split for training the universal head
universal_val.jsonl
10% split for… See the full description on the dataset page: https://huggingface.co/datasets/Dennishuang85/scent-engine-annotations.