gatineausoftware / ludwig_sql

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

build mysql image:

docker build -t mysql_ml .

run image:

docker run --name mysql -v /Users/benmackenzie/projects/mysql:/var/lib/mysql-files/ -e MYSQL_ROOT_PASSWORD=swamp -p 3306:3306 mysql_ml

so far:

  • load titanic file into mysql
  • read into a dataframe
  • define ludwig Model
  • train with data frame

to do:

  • build model definition automatically
  • provide visual mechanism to select target
  • automatically load into fastscore [generate schema, generate model conformance python, generate tar file, use fastscore deploy api to add model]

About


Languages

Language:Jupyter Notebook 77.5%Language:Python 22.5%Language:Dockerfile 0.1%