StevenHH2000 released this training dataset on March 19, 2026 for a CVPR 2026 paper on taxonomy-aware representation alignment. It consists of randomly sampled one-shot examples per category from the iNaturalist2021 dataset. The data includes images paired with text questions and coarse-to-fine category labels.
Use Cases
- Training models for few-shot hierarchical classification based on coarse-to-fine labels.
- Benchmarking taxonomy-aware representation alignment methods based on paired image-text data.
- Developing visual question-answering systems based on the 'problem' field.
- Evaluating large multimodal models on naturalist image recognition tasks.
Strengths
- Data is officially released for a CVPR 2026 conference paper.
- Examples are sampled per category from the established iNaturalist2021 dataset.
- Data fields include multimodal image-text pairs with hierarchical labels.
Limitations
- Column-level documentation is absent; field semantics must be inferred after download.
- Row count is unknown, which may limit suitability assessment.
- Description metadata is limited; actual data quality requires manual inspection after download.
Provenance
- Source
- iNaturalist2021 dataset
- Collection Method
- Randomly sampled 1-shot data per category.
- Freshness
- Last updated 2026-03-19 10:42:58; freshness should be verified.