Inference

An architecture for accelerated large-scale inference of transformer-based language models