We recommend using the mlverse/docker image which you can retrieve and run as follows:
docker pull mlverse/mlverse-base:version-0.2.0
docker run --gpus all -p 8787:8787 -d mlverse/mlverse-base:version-0.2.0
You can then connect to RStudio Server under port https://public-address:8787, followed by installing and runninng AlexNet:
remotes::install_github("r-tensorflow/alexnet")
alexnet_train::alexnet_train()