Datasets:

---

hiker-lw

/

VL-Compositionality-Benchmarks

like 0

This PR improves the dataset card for VL-Compositionality-Benchmarks by:

• Adding the zero-shot-image-classification task category to the metadata.
• Linking the dataset to the associated paper: "Cross-Modal Masked Compositional Concept Modeling for Enhancing Visio-Linguistic Compositionality".
• Providing a link to the official GitHub repository.
• Describing the various compositional benchmarks contained in this collection.
• Including important data preparation notes regarding the concatenation of split files for VL-Checklist.