Fetching the paper…

Training Tensor Attention Efficiently: From Cubic to Almost Linear Time · Around