In this competition you are tasked with transcribing lines of Deseret Alphabet images into unicode representations. The data is provided as images of various sizes that capture the content of a single line in a manuscript. You will develop a model that will transcribe each line in the **test/images/** directory and creating a submission csv file in the style of **sample_submission.csv**.

## Files

*   **train/images/** Contains transcribed png files of Deseret lines 
*   **train/labels/** Contains the corresponding labels to the ids of the train images. ex: **train/images/000cf4502.png** has transcription in **train/labels/000cf4502.txt**: 𐐮𐑌 𐐸𐐮𐑆 𐐿𐐮𐑍𐐼𐐲𐑋.
*   **test/images/** Contains unlabeled png files for your model to transcribe into Deseret to submit to the competition
*   **sample_submission.csv** - a sample submission file in the correct format

The images for this competition are courtesy of the Church History Library of the Church of Jesus Christ of Latter-day Saints, [CR 244 415](https://catalog.churchofjesuschrist.org/record/ab7480c3-3a13-4e80-8d2f-8d140caf78bd/0?view=browse&lang=eng).

The labels for this competition are courtesy of: Illinois Deseret Consortium. 2025. _Deseret (Unicode) transcription of the New Testament_. Accessed November 4, 2025, from http://faculty.las.illinois.edu/rshosted/deseret.html.