Fetching the paper…

Approximation Rate of the Transformer Architecture for Sequence Modeling · Around