Monolith is a deep learning framework for large scale recommendation modeling. It introduces two important features which are crucial for advanced recommendation system:

• collisionless embedding tables guarantees unique represeantion for different id features
• real time training captures the latest hotspots and help users to discover new intersts rapidly

Monolith is built on the top of TensorFlow and supports batch/real-time training and serving.

• Python 3.8.6
• Bazel Version 4.0.0
• Java: 1.8.0
• Kafka: 2.13-2.8.1

asdf local python 3.8.6 python -m venv .venv . .venv/bin/activate pip install -r requirements.txt 

export BAZEL_VERSION=4.0.0 curl -fLO "https://github.com/bazelbuild/bazel/releases/download/$BAZEL_VERSION/bazel-$BAZEL_VERSION-installer-darwin-x86_64.sh" chmod +x "bazel-$BAZEL_VERSION-installer-darwin-x86_64.sh" ./bazel-$BAZEL_VERSION-installer-darwin-x86_64.sh --user export PATH="$PATH:$HOME/bin" 

1. native_training:demo: bazel run //monolith/native_training:demo --output_filter=IGNORE_LOGS ⚠️ (segmentation fault)
2. batch: bazel run //markdown/demo:demo_local_runner -- --training_type=batch ⚠️ (segmentation fault)
3. streaming: ❌ static assertion failed due to requirement
4. Kubernetes:
   1. EKS: todo
   2. GKE: todo

There are a tutorial in markdown/demo on how to run distributed async training, and few guides on how to use the MonolithModel API here.

When you encounter a problem and want to clean up the bazel cache, you can use the following command:

bazel clean --expunge