zuhri025/Urdu-Munch-Curriculum-easy

Urdu Munch Curriculum Easy Filtered Urdu TTS training subset for easier, shorter audio sequences. Split train: 242549 rows Curriculum Rule audio_token_len < 120 Fields id transcript voice text timestamp duration audio_content_token_indices audio_global_embedding audio_token_len Notes This dataset was created by filtering zuhri025/Urdu-Munch-Processed-s-merged for curriculum learning.