ULSeq-TA: Ultra-Long Sequence Attention Fusion Transformer Accelerator Supporting Grouped Sparse Softmax and Dual-Path Sparse LayerNorm

Jingyu Wang, Lu Zhang, Xueqing Li, Huazhong Yang, Yongpan Liu. ULSeq-TA: Ultra-Long Sequence Attention Fusion Transformer Accelerator Supporting Grouped Sparse Softmax and Dual-Path Sparse LayerNorm. IEEE Trans. on CAD of Integrated Circuits and Systems, 43(3):892-905, March 2024. [doi]

Authors

Jingyu Wang

This author has not been identified. Look up 'Jingyu Wang' in Google

Lu Zhang

This author has not been identified. Look up 'Lu Zhang' in Google

Xueqing Li

This author has not been identified. Look up 'Xueqing Li' in Google

Huazhong Yang

This author has not been identified. Look up 'Huazhong Yang' in Google

Yongpan Liu

This author has not been identified. Look up 'Yongpan Liu' in Google