Sign in to view source links and access this dataset
Description
134,281 wildlife images across 122 species and subspecies labels, compiled by Horama. The dataset was last updated on 2026-04-29 and aggregates images from sources including iNaturalist, Wikimedia, Wikipedia, Flickr, and DuckDuckGo.
Use Cases
Train image classifiers for wildlife species based on the 122 provided labels.
Benchmark multi-source image aggregation techniques using data from iNaturalist, Wikimedia, and Flickr.
Fine-tune vision models on a dataset with a mix of wild species and domestic categories.
Develop data loaders for Hugging Face's Image feature format, decoding directly to PIL.Image.
Strengths
Contains 134,281 images, providing a substantial volume for model training.
Covers 122 distinct labels for species, subspecies, and domestic categories.
Aggregates data from five distinct public sources (iNaturalist, Wikimedia, Wikipedia, Flickr, ddg).
Limitations
Column-level documentation is absent; field semantics must be inferred after download.
The split sizes for validation and test sets are not fully specified in the provided description.
Data may reflect geographic, temporal, or source bias inherent to the contributing platforms.
Provenance
Source
Horama
Collection Method
Aggregated from multiple public online sources: iNaturalist, Wikimedia, Wikipedia, Flickr, DuckDuckGo.
Freshness
Last updated 2026-04-29 12:28:49.
License is unknown; terms of use should be verified before commercial application.