GPU Memory Usage Optimization for Backward Propagation in Deep Network Training

Re-assign community