Fetching the paper…

Sharp Models on Dull Hardware: Fast and Accurate Neural Machine Translation Decoding on the CPU · Around