I'm trying the deep dream example and I have it running on tensorflow/CPU backend.
I'm on Debian, I'm having a bad time with CUDA, I'm generally happier with plaidml running on OpenCL.
If I change the start of the code to:
library(keras) library(tensorflow) use_backend("plaidml") library(purrr)
Then I get the following error at the following lines:
> model <- application_incepti .... [TRUNCATED] INFO:plaidml:Opening device "geforce_gtx_960.0" Error in py_call_impl(callable, dots$args, dots$keywords) : KeyError: 'plaidml'