polyaxon/polyaxon

A platform for reproducible and scalable machine learning and deep learning on kubernetes


repo name	polyaxon/polyaxon
repo link	https://github.com/polyaxon/polyaxon
homepage	https://polyaxon.com
language	Python
size (curr.)	58067 kB
stars (curr.)	2350
created	2016-12-26
license	Apache License 2.0

Polyaxon

Welcome to Polyaxon, a platform for building, training, and monitoring large scale deep learning applications.

Polyaxon deploys into any data center, cloud provider, or can be hosted and managed by Polyaxon, and it supports all the major deep learning frameworks such as Tensorflow, MXNet, Caffe, Torch, etc.

Polyaxon makes it faster, easier, and more efficient to develop deep learning applications by managing workloads with smart container and node management. And it turns GPU servers into shared, self-service resources for your team or organization.

Install

TL;DR;

Create a deployment

# Create a namespace
$ kubectl create namespace polyaxon

# Add Polyaxon charts repo
$ helm repo add polyaxon https://charts.polyaxon.com

# Deploy Polyaxon
$ helm install polyaxon/polyaxon \
    --name=polyaxon \
    --namespace=polyaxon \
    -f config.yaml

Install CLI

# Install Polyaxon CLI
$ pip install -U polyaxon-cli

# Config Polyaxon CLI
$ polyaxon config ...

# Login to your account
$ polyaxon login

Please check polyaxon installation guide

If you are looking for the documentation of Polyaxon v0.4, please go here.

Quick start

TL;DR;

Start a project

# Create a project
$ polyaxon project create --name=quick-start --description='Polyaxon quick start.'

# Clone the quick start repo
$ git clone https://github.com/polyaxon/polyaxon-quick-start.git
$ cd polyaxon-quick-start

# initialize it
$ polyaxon init quick-start

Train and track logs & resources

# Upload code and start experiments
$ polyaxon run -u

# Stream logs
$ polyaxon experiment -xp 1 logs

# Stream resources
$ polyaxon experiment -xp 1 resources

Dashboard

# Start Polyaxon dashboard
$ polyaxon dashboard

Dashboard page will now open in your browser. Continue? [Y/n]: y

Notebook

# Start Jupyter notebook for your project
$ polyaxon notebook start -f polyaxonfile_notebook.yml

Notebook is being deployed for project `quick-start`

It may take some time before you can access the notebook..

Your notebook will be available on:

    http://52.226.37.54:80/notebook/root/quick-start

Tensorboard

# Start TensorBoard
$ polyaxon tensorboard start

Tensorboard is being deployed for project `quick-start`

It may take some time before you can access tensorboard.

Your tensorboard will be available on:

    http://52.226.37.54:80/tensorboard/root/quick-start

Please check our quick start guide to start training your first experiment.

Distributed training

Polyaxon supports and simplifies distributed training. Depending on the framework you are using, you need to adapt your code to enable the distributed training, and update your polyaxonfile with an environment section.

Here are some examples of using distributed training:

Hyperparameters tuning

Polyaxon has a concept for suggesting hyperparameters and managing their results very similar to Google Vizier called experiment groups. An experiment group in Polyaxon defines a search algorithm, a search space, and a model to train.