Improving Neural Network Quantization without Retraining using Outlier Channel Splitting

Open in new window