Skip to content

Latest commit

 

History

History
79 lines (65 loc) · 2.06 KB

README.md

File metadata and controls

79 lines (65 loc) · 2.06 KB

mlflow-k8s

An easy-to-use MLflow platform to track experiments, flexible more users.

Blogged (Japanese): https://cyberagent.ai/blog/research/15272/

mlflow-k8s

Prerequisites

Software Install (Mac)
Python 3.8.10 pyenv install 3.8.10
Poetry 1.1.* curl -sSL https://raw.githubusercontent.com/python-poetry/poetry/master/get-poetry.py | python
direnv brew install direnv
jq brew install jq
Docker brew cask install docker
Google Cloud SDK https://cloud.google.com/sdk/docs/quickstart-macos/
Terraform brew tap hashicorp/tap && brew install hashicorp/tap/terraform

Get Started

  1. Declare the environment variables
cp .env.example .env
vi .env
=====================
GCP_PROJECT=YOUR_GCP_PROJECT
GOOGLE_APPLICATION_CREDENTIALS=/path/to/gcp-terraform-credentials.json
TF_GCS_BUCKET=YOUR_PREFIX-tfstate
MLFLOW_GCS_BUCKET=YOUR_PREFIX-mlflow-k8s
MLFLOW_DOMAIN=YOUR_MLFLOW_DOMAIN
MLFLOW_TRACKING_URI=https://${MLFLOW_DOMAIN}/user1/
SERVICE_ACCOUNT_EMAIL=mlflow-k8s@${GCP_PROJECT}.iam.gserviceaccount.com
direnv allow .
  1. Replace the variables in .tf file for you
make tf
  1. Initialize the bucket for terraform
make bucket
  1. Run terraform
cd infra
terraform init
terraform apply
  1. Check your launched server accessible via Cloud IAP
cd ..
make access