Efficient Neural Compression with Inference-time Decoding

C. Metz, O. Bichler, A. Dupret

This paper explores the combination of neural network quantization and entropy coding for memory footprint minimization. Edge deployment of quantized models is hampered by the hars…