mosaicml / llm-foundry

LLM training code for Databricks foundation models
https://www.databricks.com/blog/introducing-dbrx-new-state-art-open-llm
Apache License 2.0
3.85k stars 504 forks source link

TransformerEngine Image Build #1204

Closed mvpatel2000 closed 1 month ago

mvpatel2000 commented 1 month ago

Adds TransformerEngine to dockerfile. This dramatically speeds up FP8 workloads