Skip to main content

Deploy your model in one line of code

Project description

Deploy your ML pipelines effortlessly, scalably and reliably

Datarmada aims at removing all the friction that comes with Machine Learning in production. We understand that Data Scientists are not trained to do that, and sometimes they are not even attracted by this Software Engineering / DevOps aspect.

This package aims at deploying your machine learning pipeline on a server in one line.

Your pipeline is deployed on an OVH server so that you own your data and it is compliant with European regulations.

Installation

Install the package python using pip

pip install auto-mlops

Deploy your pipeline

Import the Deployer class from the package.

from auto_mlops import Deployer
deployer = Deployer()

Now, deploy your pipeline by passing to the deploy method a list containing all of its elements. The pipeline elements (except for the last one) must be either :

  • A function returning transformed data if your pipeline element doesn't need to be fitted
  • An instance of a class implementing a transform method

The last element of the pipeline must be an instance of a class a predict methods, such as a scikit-learn or a Keras model.

from sklearn.linear_model import LogisticRegression

def preprocess(raw_data):
    # preprocess the data
    return preprocessed_data

class Featurizer:
    def transform(self, preprocessed_data):
        # transform the data
        return featurized_data

log_reg = LogisticRegression()
log_reg.fit(featurized_data, y)

deployer.deploy([preprocess, featurizer, log_reg])

Remember your elements must be fitted if they need to !

You will be asked for your email address so that we can keep track of the ownership of the pipelines deployed, and give you access to monitoring functions in the future.

deployer.deploy([preprocess, featurizer, log_reg])

>> Please enter your email address so that we can keep track of your pipelines:
you@example.com

>> Your pipeline has been deployed to https://cloud.datarmada.com/id

You can access your route whenever you want through deployer.route

Make predictions

You can now send data to the route by making a POST request as following

import requests

res = requests.post(
  "https://cloud.datarmada.com/id",
  json = {
    "data": your_raw_data
  }
)

print(res.json())

>> { "prediction" : prediction }

It may be possible that one of the package you are using is not available in the environment we are deploying your model. If you receive an error saying so, please email us at contact@datarmada.com so that we can fix it.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Files for auto-mlops, version 0.1.12
Filename, size File type Python version Upload date Hashes
Filename, size auto_mlops-0.1.12-py3-none-any.whl (4.0 kB) File type Wheel Python version py3 Upload date Hashes View
Filename, size auto-mlops-0.1.12.tar.gz (3.5 kB) File type Source Python version None Upload date Hashes View

Supported by

Pingdom Pingdom Monitoring Google Google Object Storage and Download Analytics Sentry Sentry Error logging AWS AWS Cloud computing DataDog DataDog Monitoring Fastly Fastly CDN DigiCert DigiCert EV certificate StatusPage StatusPage Status page