Models for the Real World

Putting the power of deep learning into practice.

Every model you build with Stream.ML comes with its own accuracy report and can be put into production immediately.
A simple API call returns the prediction values ready to be embedded directly into your application. Monitor the performance of all deployed models from your Stream.ML account.

Deployable Models

At this time we are hosting models built with our own Stream AutoML, and Google AutoML Vision, and we have plans to host a variety of popular machine learning model formats so data scientists everywhere will have the opportunity to monetize their models on a per use basis.

We are starting with TensorFlow, Keras, and PyTorch but we would be happy to host any type of machine learning model.  Please contact us if you have a model you want to monetize that we haven’t implemented yet.

Cloud Hosting

The core of our offering is immediate cloud hosting of any model you create or purchase from our marketplace.  The process is seamless with no need to worry about virtual machines, patches, updates, or up-time.  We allow for everything from development level deployments to production in a single button click.

Once hosted, you have access to a simple API and API keys that allows immediate access and tracking of all usage.

Edge Deployment

If you are making an application that would be better suited to an embedded model, we are working hard to release an Edge SDK for both iOS and Android.  As we are still developing this SDK we would be happy to hear about what features you need for this to be successful.

On-Premise Deployment

If you need more control or are making a vast number of predictions, it will make sense to host your machine learning model in your own infrastructure.  We provide everything required to quickly and easily setup your own on-premise predictions.