Has anyone experimented or installed either Docker or Anaconda on a Lamba machine? Concerned about not stepping on the installed packages.
1 Like
Yes we run training runs inside a docker container on our lambda boxes. It works well. If you want to try out a pre-built container for keras and tensorflow checkout my repo:
I just released these today so I’m looking for feedback - try them out and let me know if you run into problems.
2 Likes
Thanks very much – will have a look straight away.
what version of cuda are you running on the host? The install script will try to upgrade you to the latest driver. So far this has been fine on my machines (Titan-X’s and 1080 Ti’s).
If you don’t want to upgrade your driver, let me know which version of cuda you are running and I’ll create a container that will work with your version of cuda on the host.
1 Like
Thank you very much.
ubuntu@lambda-quad:~$ echo $PATH | grep -i cuda
/home/ubuntu/bin:/home/ubuntu/.local/bin:/usr/local/caffe-py3/build/tools:/usr/local/cuda/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin
ubuntu@lambda-quad:/usr/local/cuda$ cat /usr/local/cuda/version.txt
CUDA Version 9.0.176
that should be fine.