Add data descriptions

README.md

@@ -19,4 +19,55 @@ configs:
 license: apache-2.0
 language:
 - en
----
+---
+# Dataset Card for One Piece Image Dataset
+
+## Metadata
+
+* **Identifier:** `OnePiece_Characters`
+* **Contributions:** Mohd Hozaifa Khan
+* **License:** apache-2.0
+
+## Dataset Description
+
+This dataset contains images of famous One Piece characters scraped from the web using simple image search. It includes approximately 15 images per character for a total of 20 characters. 
+
+## Intended Use
+
+This dataset is intended for educational and research purposes, including:
+
+* **Image classification:** Train models to recognize different One Piece characters.
+* **Image generation:** Fine-tune text-to-image models to generate images of One Piece characters.
+
+## Format
+
+The dataset is organized as follows:
+
+* **Data Folder:** `one_piece_images`
+    * `train`: Contains all the images, named as `character_name_index.jpg`.
+* **Metadata File:** `metadata.jsonl`
+    * Contains metadata for each image, with each line being a JSON object containing:
+        * `file_name`: The name of the image file. (type: string)
+        * `text`: A caption or description of the image. (type: string)
+
+## Features
+
+* **image (jpg):** The image of the One Piece character. (expected type: PIL Image, JPG format)
+* **text (str):** A caption or description of the image. (expected type: string)
+
+## Slices
+
+The dataset is split into a single training split (`train`).
+
+## Acknowledgements
+
+This dataset uses images primarily from Google Images. Used `simple_image_download` for downloading images using character queries.
+
+## BibTeX citation
+
+```bibtex
+@misc{OnePiece_Characters,
+  title={One Piece Characters Image Dataset},
+  author={khanitachi},
+  year={2024}
+}