example_1k.parquet: a small sample for trying out the Precalculated Embedding Exploration app.
It contains BioCLIP 2 embeddings for 1,030 randomly sampled images from TreeOfLife-200M (embeddings only, not the images themselves). Taxonomic information and other metadata comes from catalog.parquet in the TOL-200M repo.
uuid: string
emb: list<element: float>
source_dataset: string
source_id: string
kingdom: string
phylum: string
class: string
order: string
family: string
genus: string
species: string
scientific_name: string
common_name: string
resolution_status: string
publisher: string
basisOfRecord: string
identifier: string
img_type: string