Context This dataset is created for the deep learning based Devanagari character recognition research evaluation, 2015. Content The dataset contains mixed categories for Devanagari numerals (10 classes) and consonants (36 classes). Dataset is explicitly separated into train and test set. **Train set contains total 78,200** samples with 1700 samples per class for total 46 classes and **test set contains total 13,800** samples with 300 samples per class for total 46 classes. Dataset is collected from the school level students. This dataset is collected and maintained by the following research members, 1. Shailesh Acharya 2. [Ashok Kumar Pant][1] 3. Prashnna Kumar Gyawali Citation Please cite in your publications if it helps your research: @inproceedings{ashok2015deep, author={S. Acharya and A. K. Pant and P. K. Gyawali}, booktitle={2015 9th International Conference on Software, Knowledge, Information Management and Applications (SKIMA)}, title={Deep learning based large scale handwritten Devanagari character recognition}, year={2015}, pages={1-6}, month={Dec} } [1]: http://ashokpant.github.io/