Update README.md

README.md

@@ -34,3 +34,17 @@ configs:
   - split: mtinstruct_en2th
     path: data/mtinstruct_en2th-*
 ---
+# Thai Aligninstruct Dataset
+
+This project was try to make English-Thai Aligninstruct.
+
+I use simalign for making the High Quality Word Alignments Without Parallel Training Data Using Static and Contextualized Embeddings and use [deepcut](https://github.com/rkcosmos/deepcut) (by [lekcut](https://github.com/PyThaiNLP/LEKCut)) to tokenize Thai word.
+
+The data use [scb-mt-en-th-2020](https://huggingface.co/datasets/scb_mt_enth_2020).
+
+
+## References
+
+- Aligninstruct: [https://arxiv.org/abs/2401.05811v1](https://arxiv.org/abs/2401.05811v1)
+- SimAlign: High Quality Word Alignments Without Parallel Training Data Using Static and Contextualized Embeddings: [https://aclanthology.org/2020.findings-emnlp.147/](https://aclanthology.org/2020.findings-emnlp.147/)
+- scb-mt-en-th-2020: A Large English-Thai Parallel Corpus: [https://arxiv.org/abs/2007.03541](https://arxiv.org/abs/2007.03541)