Google Cloud ML Tensorflow Version

Question

The docs for setting up Google Cloud ML suggest installing Tensorflow version r0.11. I've observed that TensorFlow functions newly available in r0.12 raise exceptions when run on Cloud ML. Is there a timeline for Cloud ML supporting r0.12? Will switching between r0.11 and r0.12 be optional or mandatory?

Carleton

Answer 1

Yes, you can specify --runtime-version=0.12 to get a 0.12 build. This is a new feature and is documented at https://cloud.google.com/ml/docs/concepts/runtime-version-list

Note, however, that the 0.12 build is not yet considered stable and the exact Tensorflow build provided may change. Once the 1.0 version of Tensorflow is available, that will also be supported and the pre-1.0 versions of Tensorflow will begin to be deprecated.

(See https://cloud.google.com/sdk/gcloud/reference/beta/ml/jobs/submit/training for usage.)

Melchor

I saw it's not possible to specify a minor version. Is there any way to use the 0.12.1 build, instead of 0.12.head? The problem is that 0.11.0 no longer authenticates with GCS (for local development), so we have to use 0.12.1 locally. But we've been unable to get that same fixed version deployed on Cloud ML, which is causing a lot of headaches for us. – Longitude

The 0.12 version is actually a 0.12.1 build. (In particular, it is this build referenced on the download tensorflow page: storage.googleapis.com/tensorflow/linux/cpu/…) Prior to 1.0, we don't plan on supporting every minor TF version, as there have been many breaking changes. – Melchor