Go to website
AI and Data
Our AI Algorithms and Data Usage
Where can I find the Tarteel dataset?
The Tarteel dataset can be found at https://tarteel.ai/dataset Currently we only have one version of the dataset, however, we plan on releasing more data as we clean it up and improve the quality of the audio and labels. We recommend taking a look at the following repos for technical resources on using the Tarteel Dataset. tarekeldeeb/DeepSpeech-Quran [tarteel-io/tarteel-ml](https://github.com/tarteel-io/tart
Don't poor reciters/recitations make your algorithm perform worse?
Collecting both good and bad data is helpful because it allows our algorithm to learn what is a good and bad recitation. We have a mix of automatic & manual evaluation to identify whether a verse was recited correctly or not.
What ML model/algorithms/techniques do you use?
Our speech-to-text model is primarily based off of Baidu's DeepSpeech2 (Original Paper). However, it has been modified and improved to meet our use case.
Do you have an API I can use?
We do not have a public API for other developers/applications to use and do not plan to support one at this time. If you are looking for digital Quran resources, we recommend Quran.com, Quranenc.com, and alquran.cloud.