Exploring district-level dialect identification in India using the Vaani dataset and various pre-trained speech models.