Skip to content

Easy to deploy MLFlow(machine learning lifecycle system)

Notifications You must be signed in to change notification settings

4pygmalion/mlflow

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

20 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

mlflow

machine learning lifecycle sys

Install

  • GCP 이용시 (invalid Host header 오류발생) -> 최신버전 도커설치 필요
$ sudo snap refresh docker --channel=latest/edge
  1. Docker official image pull
$ docker pull ghcr.io/mlflow/mlflow
  1. DB setting
$ export MYSQL_DATABASE=<db_name>
$ export MYSQL_USER=<user>
$ export MYSQL_PASSWORD=<password>
$ export MYSQL_ROOT_PASSWORD=<root_password>
$ export MYSQL_DIR=<host_mysql_dir>

// example
export MYSQL_DATABASE=mlflow
export MYSQL_USER=mlflow
export MYSQL_PASSWORD=mlflow
export MYSQL_ROOT_PASSWORD=mlflow
export MYSQL_DIR=/home/hoheon/repositories/mlflow_db
  1. MLflow trackign server setting
$ export MLFLOW_PORT=<mlflow_port>
$ export ARTIFACT_DIR=<artifact_dir>

//exmaple
export MLFLOW_PORT=5000
export ARTIFACT_DIR=/home/hoheon/repositories/mlflow
  1. Run
// rootless 도커사용시 docker compose
$ docker-compose pull
$ docker-compose build --pull
$ docker-compose up --build --remove-orphans --detach # build after pull latest

With sudo

$ sudo -E docker-compose pull (-E option: 환경변수 유지)
$ sudo -E docker-compose build --pull
$ sudo -E docker-compose up --build --remove-orphans --detach

rootless docker을 이용한 배포

$ bash run.sh

Contents