Chimera: efficiently training large-scale neural networks with bidirectional pipelines

Shigang Li 0002, Torsten Hoefler. Chimera: efficiently training large-scale neural networks with bidirectional pipelines. In Bronis R. de Supinski, Mary W. Hall, Todd Gamblin, editors, SC '21: The International Conference for High Performance Computing, Networking, Storage and Analysis, St. Louis, Missouri, USA, November 14 - 19, 2021. pages 27, ACM, 2021. [doi]

Authors

Shigang Li 0002

This author has not been identified. Look up 'Shigang Li 0002' in Google

Torsten Hoefler

This author has not been identified. Look up 'Torsten Hoefler' in Google