This PR adds a comprehensive dataset card for the V-PETL Bench, including its purpose, relevant task categories, and essential links.
Specifically, it:
- Adds `task_categories` (image-classification, video-classification, object-detection, image-segmentation) to reflect the benchmark's scope.
- Includes the `cc-by-nc-4.0` license.
- Links to the official paper on Hugging Face papers: https://huggingface.co/papers/2402.02242.
- Links to the project homepage: https://v-petl-bench.github.io/.
- Provides an introduction to the benchmark, details on data preparation, usage, and citation.
- Incorporates relevant images from the GitHub repository for better visual representation.
