Update README.md

README.md

@@ -34,7 +34,9 @@ language:
 - en
 - km
 ---
-This dataset contains image URLs paired with human-readable captions in both English and Khmer. The captions are extracted from the `visheratin/laion-coco-nllb` dataset and filtered for quality and language accuracy using the score metric. Only entries with valid and reachable image URLs and high-confidence caption pairs were retained.
+## Description
+
+This dataset contains image URLs paired with human-readable captions in both English and Khmer. The captions are extracted from the [`visheratin/laion-coco-nllb`](https://huggingface.co/datasets/visheratin/laion-coco-nllb) dataset and filtered for quality and language accuracy using the score metric. Only entries with valid and reachable image URLs and high-confidence caption pairs were retained.
 Each example contains:
 - `id`: Unique ID for the sample
 - `url`: Direct image URL
@@ -42,7 +44,7 @@ Each example contains:
 - `km_caption`: Khmer translation of the caption
 - `score`: Confidence score (e.g. NLLB translation quality or cosine similarity)
 
-### Use Cases
+## Use Cases
 - Multimodal Machine Translation (image ↔ text in multiple languages)
 - Khmer-English captioning tasks
 - Low-resource language training for image-to-text or text-to-image models