A distributed hyperparameter optimizer for machine learning that lives in Docker
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
docker
extra
hyperdock
tests
web
.coveragerc
.dockerignore
.gitignore
.travis.yml
LICENSE
Pipfile
Pipfile.lock
README.md
new_release.sh
requirements.txt
setup.py

README.md

Hyperdock

Hyperdock logo

PyPI Docker Pulls Travis Codecov

A simple framework for distributed hyperparameter optimization in Docker.

Philosophy

The main idea of Hyperdock is universal hyperparameter optimization that can optimize a target written in any language or using any framework as long as it can be deployed in a Docker image.

Hyperdock diagram

From the Hyperdock WebUI the user can add trials that specify the target image and the hyperparameter space.

The Hyperdock Supervisor then reads the user-defined trial from the Mongo database and creates sets of parameter combination to be evaluated. These jobs are added to the work queue.

The Hyperdock Workers dequeues jobs from the work queue and then evaluates the target image with these parameters. During this the Hyperdock Supervisor monitors the progress of the jobs and the status of the workers.

The Hyperdock Supervisor, Workers and WebUI need not run on the same host. They only need to be able to access the same Mongo database and the workers need to have access to any data required by the target image.

The target image receives the parameters from the json file /hyperdock/params.json. Once the target image has evaluated the parameters it simply writes the loss to file /hyperdock/loss.json with the option of storing important files to /hyperdock/out. Logs from the target image are periodically tailed from the workers to the WebUI.

Running

Setting up the Hyperdock system can seem a bit complicated but once it is up it quite easy to use.

You can either use the built Docker images for Hyperdock or run the parts directly on the host.

Supervisor

To start the Hyperdock Supervisor using the Docker image run the following command:

docker run -it --rm --name hyperdock-supervisor --link hyperdock-mongo \
  erikgartner/hyperdock-supervisor:latest \
  --mongodb mongodb://hyperdock-mongo:27017/hyperdock

Or run it on your host with Python 3.6 and install with pip:

pip install hyperdock
hyperdock-supervisor --mongodb mongodb://localhost:27017/hyperdock
Options:
  • --mongo mongodb://localhost:27017/hyperdock URL to the Mongo database

For full arguments to the supervisor run: hyperdock-supervisor --help.

Worker

To start the Hyperdock Worker using the Docker image run the following command:

docker run -it --rm -v /var/run/docker.sock:/var/run/docker.sock \
  --link hyperdock-mongo \
  -v $(pwd):$(pwd) \
  erikgartner/hyperdock-worker:latest \
  --mongodb mongodb://hyperdock-mongo:27017/hyperdock
Options:
  • -v $(pwd):$(pwd) mirrors the path structure from the host in to the Docker container. This is needed since the paths must be the the same when the worker starts the Target Image and mounts the data and results folders.
  • -v /var/run/docker.sock:/var/run/docker.sock gives the Docker image access to control the outer Docker daemon. This is crucial for worker to start new containers

Or run it on your host with Python 3.6 and install with pip:

pip install hyperdock
hyperdock-worker --mongodb mongodb://localhost:27017/hyperdock

For full arguments to the worker run: hyperdock-worker --help.

Note: That since the Hyperdock Worker needs to control Docker and access files on the host computer it might be easier to run it directly on the host.

WebUI

To start the Hyperdock WebUI using the Docker image run the following command:

docker run -it --rm --name hyperdock-webui --link hyperdock-mongo \
  -e ROOT_URL=http://localhost:3000/ \
  -e MONGO_URL=mongodb://hyperdock-mongo:27017/hyperdock \
  -p 3000:3000 \
  erikgartner/hyperdock-webui:latest
Options:
  • -e MONGO_URL=mongodb://hyperdock-mongo:27017/hyperdock sets the Mongo database
  • -p 3000:3000 publish the port to the host

Or run it on your host with Meteor:

# Setup Meteor
curl https://install.meteor.com/ | sh # Install Meteor
cd web/
meteor npm install

# Start WebUI
export MONGO_URL=mongodb://localhost:27017/hyperdock
meteor run

Target Image

For the Target Image the following volumes are mounted:

  • /hyperdock/
    • loss.json write the final loss here
    • params.json contains the parameters for this run
    • graphs.json optional file that contains graphs that Hyperdock will plot; format described here
    • out/ use this to write any other files to the result folder
  • /data a read only folder that contains any external data needed

See the Dockfile template for an example. It is available as a demo image named erikgartner/hyperdock-demo:latest. By default it outputs 0 as its loss but by setting the environment variable FUNCTION to a python expression (for example a + b) you can compute an arbitrary loss based on the Hyperdock parameters.

Mongo database

To start a Mongo database you can use this simple Docker command or use any normal Mongo instance.

# Starts mongo db, add --bind_ip_all to listen on all interfaces.
docker run --name hyperdock-mongo -p 27017:27017 -d mongo

Building

If you want to develop Hyperdock it is recommended that you use Pipenv to manage the Python version the package dependencies.

The WebUI is built using Meteor which needs to be installed prior to development.

License

Copyright 2018 Erik Gärtner

Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at

http://www.apache.org/licenses/LICENSE-2.0