From_Shloka_to_Shobda_Corpus

The Shloka to Shobda dataset is organized as a collection of source-specific Sanskrit and Bengali text files, where each language is stored separately in plain text format. The dataset includes content collected from religious scriptures, classical literature, educational materials, tutorials, and translated documents. Each Sanskrit file has a corresponding Bengali translation file with sentence-level alignment maintained line by line. In other words, a sentence appearing at a particular line number in the Sanskrit text directly corresponds to the translated Bengali sentence at the same line number in the paired file. This structure makes the dataset suitable for neural machine translation, bilingual corpus construction, tokenizer training, and other low-resource NLP research tasks involving Sanskrit and Bengali.