Skip to content

Latest commit

 

History

History
23 lines (15 loc) · 1.67 KB

README.md

File metadata and controls

23 lines (15 loc) · 1.67 KB

Dutch Open Speech Recognition Benchmark

Welcome to the website created for researchers and developers of Dutch ASR technology to share obtained evaluation results.

It is accessible at: https://opensource-spraakherkenning-nl.github.io/ASR_NL_results

Local setup

To test changes locally, you can set up the website on your own machine by following the official Github guide.

File structure

Each page of the website has its own markdown file. index.md corresponds to the main page of the website. Each folder must contain at least the WER/main performance metric results, in the format of a matrix. Optional, but recommended, information that can be added is time performance, hardware setup, or normalization steps applied.

Each researcher/developer should create a new folder with an unique and semnificative name. The structure inside that folder can vary, according to how the researcher/developer wants to structure it. Some ideas include reporting the WER and time in 2 separate files, for all datasets/corpora, or the same structure as in the UT folder, reporting all results for each dataset/corpus used.

Add your own work!

We encourage all Dutch ASR researchers and developers to upload the results they obtained on our website. This can be simply done by:

  1. Forking this repository
  2. Adding the results and testing the changes on your fork locally
  3. Opening a PR from your fork to this repository

Feedback

If you have any suggestions or feedback to share about this website or its structure, we encourage you to open an issue or a pull request.