Block Transformer: Global-to-Local Language Modeling for Fast Inference

Open in new window