Deployment of deep learning


i have a few questions but unsure where to park them.

  1. What is the differences between paperspace core vs gradient?
  2. if i have trained a deep learning model, should i look at core or gradient to deploy it?
  3. is it possible to have a live webcam or video streaming in paperspace?
  4. is it possible to have a map overlay in paperspace?
    thank you

Hi @lchunleo

  1. Gradient is a software platform for training and deploying machine learning models. CORE provides easy access to virtual-machines and can be used for machine learning but also many other applications from high-end 3D graphics to scientific computing.
  2. Both options are valid. Gradient provides a higher-level of abstraction where you give us a model, a container and some parameters and we’ll host the model on a serverless load-balanced cluster. In CORE, you would connect to your VM, download/upload the model, and then deploy/host it on that VM. CORE is great for simple deployments since the experience is just like a traditional operating system and would be very familiar. Gradient is targeted towards more advanced deployments with multiple instances running but is easy to use as well.
  3. Our Windows instances include USB redirection in the desktop app which would support webcams or you can use a third-party remote desktop app.
  4. Not sure what you mean in terms of a “map overlay” – if you could elaborate on this, that would be super helpful.


will hold back the last question as it may not be relevant now.

is paperspace for free version, allow developers to push the docker container / image for deployment? how much ram size is allocated? thank you