llm-foundry

9fcdfe5b - Add support to run MPT-1b training on Habana device (HPU) using DeepSpeed. (#527)