Datasets
Sadeed Tashkeela
Open Source
A large, high-quality Arabic diacritized corpus for training and evaluating Arabic diacritization models. The training split is built from a cleaned version of the Tashkeela corpus; the test split is a refined version of the Fadel Tashkeela test set.
Let's talk about what you're trying to build.
Tell us the problem. We'll tell you honestly whether AI is the right answer.