1 paper
Aleksei Samarin, Artem Nazarenko, Egor Kotenko +3
This paper presents a novel approach to neural network compression that addresses redundancy at both the filter and architectural levels through a unified framework grounded in inf…