Closed wyhhyw closed 7 years ago
No there isn't a way to do this.
Why not install them on the same server? On Wed, Aug 16, 2017 at 8:59 PM wyhhyw notifications@github.com wrote:
I have installed the keras backend with tensorflow on one server, and rstudio-server on the other. Then can I use keras from my rstudio-server with remote connection to the keras?
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/rstudio/keras/issues/103, or mute the thread https://github.com/notifications/unsubscribe-auth/AAGXx3oyW7kMn-9HJq1P7trhFHc9CRi9ks5sY6yFgaJpZM4O5uMy .
Because we have several servers, and we want to make different servers to do different things. And we want to use the keras remotely, thus we can use it both on the server, or personal laptop. I think it is reasonable that the keras server provides service just like rstudio-server, and we use it remotely. However, I'd like to use rstudio-server to connect to the keras anyway.
That is exactly what people use rstudio-server for: they put whatever hardware and software they need in one place then have R + a thin-client to access it. Honestly compared to what keras is doing computationally R and RStudio Server won't add any material overhead.
On Wednesday, August 16, 2017, wyhhyw notifications@github.com wrote:
Because we have several servers, and we want to make different servers to do different things. And we want to use the keras remotely, thus we can use it both on the server, or personal laptop. I think it is reasonable that the keras server provides service just like rstudio-server, and we use it remotely. However, I'd like to use rstudio-server to connect to the keras anyway.
— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/rstudio/keras/issues/103#issuecomment-322958869, or mute the thread https://github.com/notifications/unsubscribe-auth/AAGXx9dYFxc7uMhrL9WfkIbugDXS3X9xks5sY7J1gaJpZM4O5uMy .
I have installed the keras backend with tensorflow on one server, and rstudio-server on the other. Then can I use keras from my rstudio-server with remote connection to the keras?