Spoken language identification consists in recognizing a language based on a sample of speech from an unknown speaker. The traditional approach for this task mainly considers the phonothactic information of languages. However, for marginalized languages