Comment by choxi

Did you set up your own infrastructure? I’ve been looking for a Heroku-like service to deploy some of these GPU intensive models