Fetching the paper…

Soft-to-Hard Vector Quantization for End-to-End Learning Compressible Representations · Around