FAQ - Daeploy

Frequently asked questions

General

What kind of algorithms can be deployed?

Any algorithm you want! It could be anything from a machine learning model to a basic control system. In fact, it could be any piece of source code you want. It is up to you, and that's the beauty of Daeploy!

Is Daeploy a MLOPS tool?

No, Daeploy enters the project when a trained model already exists. Daeploy creates an application around the trained model and makes the model a part of the production-ready application.

Which ML frameworks does Daeploy support?

Daeploy supports all major ML frameworks, for instance

PyTorch
TensorFlow
Keras

And many more!

I found a bug! How can I report it?

First of all, thank you for wanting to report the bug and improving Daeploy!

We have a public issue tracker on GitHub, please follow the instructions here.

I'm interested seeing the roadmap for Daeploy, is there a place where I can follow the development?

First of all, thank you for showing interest in Daeploy!

We have a public issue tracker and a public development board on GitHub, you find it here.

Ps, for the latest news we recommend you to join our community on Slack!

Technical

Can the Daeploy Manager run on a Windows Server?

Yes! The only requirement Daeploy has is docker and access to the docker daemon. If the server can run docker, Deaploy can run!

Does the machine/server running the Daeploy Manager need internet access?

No! Daeploy works without internet access.

First, you need to pull the Daeploy Manager image from here from a computer with internet access. You then need to save the image as a tar file using docker save. The tar file can then be transferred to the wanted machine/server using SCP or a similar tool.

When the machine has the Daeploy Manager image, we recommend following our documentation for how to configure the manager for the production. You find the relevant documentation here.

Can my models be deployed to the cloud?

Yes! You can use Daeploy to on any machine or virtual machine, be it on premise or on the cloud. Our only requirement is Docker and access to Docker daemon.