Abstract page for arXiv paper 2104.04473: Efficient Large-Scale Language Model Training on GPU Clusters Using Megatron-LM...