Update README.md

README.md

@@ -47,6 +47,8 @@ size_categories:
 
 A token classification dataset for Myanmar (Burmese) chunk segmentation, formatted for sequence labeling tasks using the BIO tagging scheme.
 
+📓 **Dataset Creation Notebook**: [myanmar-text-segmentation-dataset.ipynb](https://github.com/chuuhtetnaing/myanmar-language-dataset-collection/blob/main/Myanmar%20Text%20Segmentation/myanmar-text-segmentation-dataset.ipynb)
+
 📓 **Fine-Tuning Notebook**: [myanmar-text-segmentation-fine-tuning.ipynb](https://github.com/chuuhtetnaing/myanmar-language-dataset-collection/blob/main/Myanmar%20Text%20Segmentation/myanmar-text-segmentation-fine-tuning.ipynb) (based on the [HuggingFace Token Classification Guide](https://huggingface.co/docs/transformers/en/tasks/token_classification))
 
 🚀 **Try it out**: [Myanmar Text Segmentation Demo](https://huggingface.co/spaces/chuuhtetnaing/myanmar-text-segmentation-app)
@@ -61,8 +63,6 @@ For example, the unsegmented text `မြန်မာနိုင်ငံတွ
 
 Derived from [chuuhtetnaing/myanmar-wikipedia-dataset](https://huggingface.co/datasets/chuuhtetnaing/myanmar-wikipedia-dataset).
 
-📓 **Dataset Creation Notebook**: [myanmar-text-segmentation-dataset.ipynb](https://github.com/chuuhtetnaing/myanmar-language-dataset-collection/blob/main/Myanmar%20Text%20Segmentation/myanmar-text-segmentation-dataset.ipynb)
-
 ### Processing Pipeline
 
 1. **Paragraph extraction**: Each Wikipedia article is split by newlines, preserving full paragraphs as individual rows rather than sentence-by-sentence. This design allows models to handle multi-sentence inputs without requiring line-by-line splitting at inference time.