Ron Banner

Model Compression and Hardware Acceleration for Neural Networks: A Comprehensive Survey

Scaling Up Your Kernels to 31×31: Revisiting Large Kernel Design in CNNs

Deep Learning for Generic Object Detection: A Survey

ResRep: Lossless CNN Pruning via Decoupling Remembering and Forgetting

Switchable Normalization for Learning-to-Normalize Deep Representation

Post training 4-bit quantization of convolutional networks for rapid-deployment

Scalable methods for 8-bit training of neural networks

Norm matters: efficient and accurate normalization schemes in deep networks