wangcaihua / monolith

ByteDance's Recommendation System

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Monolith(磐石)

What is it?

Monolith is a framework for training/serving large scale sparse embedding models. It is built on the top of TensorFlow 2 and estimator APIs.

Quick start

Build from source

Currently, we only support compilation on the Linux.

Frist, download bazel 3.1.0

wget https://github.com/bazelbuild/bazel/releases/download/3.1.0/bazel-3.1.0-installer-linux-x86_64.sh && \
  chmod +x bazel-3.1.0-installer-linux-x86_64.sh && \
  ./bazel-3.1.0-installer-linux-x86_64.sh && \
  rm bazel-3.1.0-installer-linux-x86_64.sh

Then, prepare a python environment

pip install -U --user pip numpy wheel packaging requests opt_einsum
pip install -U --user keras_preprocessing --no-deps

Finally, you can build any target in the monolith. For example,

bazel run //monolith/native_training:demo --output_filter=IGNORE_LOGS

About

ByteDance's Recommendation System

License:Other


Languages

Language:Python 59.5%Language:C++ 36.9%Language:Starlark 3.6%Language:Shell 0.0%