no code implementations • 11 Apr 2023 • Guangyong Wei, Zhikui Duan, Shiren Li, Guangguang Yang, Xinmei Yu, Junhua Li
With regard to the HKUST and WSJ eval92 datasets, CER and WER will be improved by 0. 3% and 0. 2%, respectively, when parameters in Sim-T are 40% less than the baseline Transformer.