We present CodeReef - an open platform to share all the compo-
nents necessary to enable cross-platform MLOps, i.e. automating
the deployment of ML models across diverse systems in the most
efficient way. We also introduce the CodeReef solution - a way to
package and share models as portable, reproducible and customiz-
able archive files. Such ML packs include JSON meta descriptions,
Python APIs, CLI actions and portable workflows necessary to au-
tomatically build, benchmark, test and customize models across
diverse platforms, AI frameworks, libraries, compilers and data
sets. We demonstrate several CodeReef solutions to automatically
build, run and measure object detection based on SSD-Mobilenets,
TensorFlow and COCO dataset from the latest MLPerf inference
benchmark across a wide range of platforms from Raspberry Pi,
Android phones and IoT devices to HPC servers with powerful
GPUs. Our long-term goal is to help the community participate in
reproducible ML benchmarking and compare the different ML/soft-
ware/hardware stacks using online CodeReef dashboards, automate
the creation of new ML benchmarks in collaboration with work-
group such as MLPerf, and share ML models as production-ready
packages along with research papers.