keras-sacred

Run sacred with keras on local machine and on your server

Preparation

• Install mongodb: apt-get install mongodb
• Install sacred: pip install sacred
• Install sacredboard locally (on Python 3 environment): pip install https://github.com/chovanecm/sacredboard/archive/develop.zip

Set up your mongodb

• Create the following directory structure wherever you want:

mkdir MongoDB
mkdir MongoDB/db MongoDB/Log

• Create the mongodb configuration file:

# mongod.conf

# Where and how to store data.
storage:
  dbPath: <path-to>/MongoDB/db
  journal:
    enabled: true

# where to write logging data.
systemLog:
  destination: file
  logAppend: true
  path: <path-to>/MongoDB/Log/mongod.log

# network interfaces
net:
  port: 27017
  bindIp: 127.0.0.1

Note: If you are running this on the server just comment out the bindIp line. Also, if the port is taken use 27018 etc.

• Start the mongodb daemon:

mongod --config <path-to>/MongoDB/mongod.conf

Setup your keras script

• Set up sacred to run with your keras script like the example given

Run sacredboard and watch your experiments

• For local experiments: sacredboard -m <db-name>
• For monitoring server experiments: sacredboard -m <server-static-ip>:<port>:<db-name>