Hugging Face's logo

Datasets:
zachz
/
Vinoground-PC

Tasks:
Video Classification
Modalities:
Text
Formats:
parquet
Languages:
English
Size:
1K - 10K
ArXiv:
Libraries:
Datasets
Dask
Polars
License:
Dataset card Data Studio Files
xet
 Community
1
Dataset Viewer
Auto-converted to Parquet Duplicate
video
video
major
stringclasses
3 values
action
action
action
action
viewpoint
viewpoint
action
action
action
action
viewpoint
viewpoint
action
action
action
action
object
object
object
object
object
object
action
action
action
action
viewpoint
viewpoint
action
action
action
action
action
action
object
object
object
object
object
object
object
object
action
action
object
object
action
action
action
action
action
action
object
object
object
object
action
action
viewpoint
viewpoint
action
action
action
action
action
action
viewpoint
viewpoint
action
action
object
object
object
object
viewpoint
viewpoint
viewpoint
viewpoint
action
action
object
object
object
object
viewpoint
viewpoint
action
action
action
action
object
object
action
action
object
object
object
object
action
action
End of preview. Expand in Data Studio

Vinoground-PC

Video pair classification dataset from Vinoground (Zhang et al., 2024).

Each row contains a video clip and its major category (action/viewpoint/object). The MTEB task generates same-category (positive) and different-category (negative) pairs to test whether embeddings capture temporal/compositional differences.

Source: HanSolo9682/Vinoground — 500 counterfactual pairs, 1000 individual clips.
Downloads last month
14
Number of rows:
1,000
Total file size:
637 MB

Paper for zachz/Vinoground-PC