datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
negativesirl
CMU Campus Negative Images Dataset
Dataset Description
This dataset contains negative example images collected around Carnegie Mellon University campus in Pittsburgh, Pennsylvania. These images represent non-lanternfly subjects that were used as negative training data for binary classification models. The dataset was created as part of an academic project (Project 1) focused on developing machine learning models for invasive species identification.
Dataset Summary… See the full description on the dataset page: https://huggingface.co/datasets/rlogh/negativesirl.generative-negative-mining-datasetThis dataset consists of 278 samples. 122 of the images have 4 variations, 139 of the images
have 3 variations, and 17 of the images have 2 variations.roof-damage-negative-evidence
Roof Damage Negative Evidence Dataset
This dataset structure documents examples where roof evidence suggests fake hail indicators, non-storm damage, poor documentation, insufficient evidence, or mixed/inconclusive conditions.
The initial public release should be a dataset shell: this dataset card, schema, CSV template, and JSONL examples. Add real images or videos only after privacy review.
Labels
fake_hail_indicator
non_storm_damage
poor_documentation… See the full description on the dataset page: https://huggingface.co/datasets/InspectorRoofing/roof-damage-negative-evidence.