Dzongkha to English translation using the neural machine translation approach

<span>In this era of technology, a communication barrier is thing the past. With each passing day, different types language-based applications are being launched. There 109 official languages Google has translated to date. However, Dzongkha translation not been studied. The purpose paper was study English translation. parallel corpus collected from development commission Bhutan. dataset consisted 53018 sentence pairs. Unique words in and were 13,393 12,506 respectively. Different neural machine models implemented. experimental results show that bleu score Seq2Seq followed fluctuating trend. transformer model increases gradually. It observed outperformed models. highest accuracy lowest training loss obtained 84.46% 0.014858 respectively with 64.89.</span>