SLAB: Efficient Transformers with Simplified Linear Attention and Progressive Re-parameterized Batch Normalization.

Jialong Guo, Xinghao Chen, Yehui Tang, Yunhe Wang

Browse the full ICML paper archive.