Fetching the paper…

Extremely Low Bit Transformer Quantization for On-Device Neural Machine Translation · Around