Improved Quantization Strategies for Managing Heavy-tailed Gradients in Distributed Learning

Open in new window