Misc insight

[Yikun]

https://databricks.com/session_na20/end-to-end-deep-learning-with-horovod-on-apache-spark
介绍uber的深度学习整体流程，结合horovod和spark，如何基于spark构建端到端的DL

https://databricks.com/session_na20/accelerating-mlflow-hyper-parameter-optimization-pipelines-with-rapids

使用MLFlow帮助AI开发者，利用hyperopt进行调参，RAPIDS加速

https://databricks.com/session_na20/how-not-to-scale-deep-learning-in-6-easy-steps
sowen简单的例子介绍scale的手段，GPU、horovod、petastorm

https://databricks.com/session_na21/enabling-vectorized-engine-in-apache-spark
IBM vectorize

https://databricks.com/session_na21/stage-level-scheduling-improving-big-data-and-ai-integration
stage level sche

HPC:
CENTOS AND HPC: IT’S OKAY, WE ARE MOVING ON
DOCKER VS SINGULARITY VS SHIFTER VS UGE CONTAINER EDITION
https://github.com/hpcng