DNN Memory Footprint Reduction via Post-Training Intra-Layer Multi-Precision Quantization

Open in new window