the fastest way to deploy machine learning models
Project description
omega|ml is the fastest way to deploy machine learning models
omega|ml takes just a single line of code to
deploy & embed machine learning models from a easy-to-use REST API
implement data pipelines quickly, without memory limitation, all from a Pandas-like API
enable scalable model training using the integrated pure-Python compute cluster
In addition you can
collaborate on data science projects easily (using Jupyter Notebook)
deploy beautiful dashboards right from your Jupyter Notebook (using dashserve)
Documentation: https://omegaml.github.io/omegaml/
Get started in < 5 minutes
Start the omega|ml server right from your laptop or virtual machine
$ wget https://raw.githubusercontent.com/omegaml/omegaml/master/docker-compose.yml
$ docker-compose up -d
Jupyter Notebook is immediately available at http://localhost:8899 (omegamlisfun to login). Any notebook you create will automatically be stored in the integrated omega|ml database, making collaboration a breeze. The REST API is available at http://localhost:5000.
Already have a Python environment (e.g. Jupyter Notebook)? Leverage the power of omega|ml by installing as follows:
# assuming you have started the server as per above
$ pip install omegaml
Examples
Get more information at https://omegaml.github.io/omegaml/
# transparently store Pandas Series and DataFrames or any Python object
om.datasets.put(df, 'stats')
om.datasets.get('stats', sales__gte=100)
# transparently store and get models
clf = LogisticRegression()
om.models.put(clf, 'forecast')
clf = om.models.get('forecast')
# run and scale models directly on the integrated Python or Spark compute cluster
om.runtime.model('forecast').fit('stats[^sales]', 'stats[sales]')
om.runtime.model('forecast').predict('stats')
om.runtime.model('forecast').gridsearch(X, Y)
# use the REST API to store and retrieve data, run predictions
requests.put('/v1/dataset/stats', json={...})
requests.get('/v1/dataset/stats?sales__gte=100')
requests.put('/v1/model/forecast', json={...})
Use Cases
omega|ml currently supports scikit-learn out of the box. Need to deploy a model from another framework? Open an issue at https://github.com/omegaml/omegaml/issues or drop us a line at support@omegaml.io
Machine Learning Deployment
deploy models to production with a single line of code
serve and use models or datasets from a REST API
Data Science Collaboration
get a fully integrated data science workplace within minutes [1]
easily share models, data, jupyter notebooks and reports with your collaborators
Centralized Data & Compute cluster
perform out-of-core computations on a pure-python or Apache Spark compute cluster [2]
have a shared NoSQL database, out of the box, that behaves like a Pandas dataframe [3]
use a compute cluster to train your models with no additional setup
Scalability and Extensibility
scale your data science work from your laptop to team to production with no code changes
integrate any machine learning framework or third party data science platform with a common API
Towards Data Science recently published an article on omega|ml: https://towardsdatascience.com/omega-ml-deploying-data-machine-learning-pipelines-the-easy-way-a3d281569666
[1] supporting scikit-learn, Spark MLLib out of the box, Keras and Tensorflow available shortly. Note the Spark integration is currently only available with the enterprise edition. [2] using Celery, Dask Distributed or Spark [3] leveraging MongoDB’s excellent aggregation framework
In addition omega|ml provides an easy-to-use extensions API to support any kind of models, compute cluster, database and data source.
Enterprise Edition
omega|ml Enterprise Edition provides security on every level and is ready made for Kubernetes deployment. It is licensed separately for on-premise, private or hybrid cloud. Sign up at https://omegaml.io
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for omegaml-0.10.2b11-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | fa7fd97852fd510ac0a52c57b7cde0273bf3ffc68a0eababbe35f7d1b5743e63 |
|
MD5 | 7dc131d75ac08ba61e9e4c8112165947 |
|
BLAKE2b-256 | abb57348b6a21121e1d6cd6d9f36275bc84e596f2d8c9e99b56043b14e4d3333 |