Keras: support TensorFlow 0.8 distributed computing

Created on 15 Apr 2016  路  4Comments  路  Source: keras-team/keras

The just released TensorFlow v0.8.0 RC0 has supported distributed computing as announced here. Can we support distributed computing by this way of tensorflow? (instead of by Spark)

stale

Most helpful comment

Yes, you can do that, by setting your own TF session via K.set_session(sess).

All 4 comments

Yes, you can do that, by setting your own TF session via K.set_session(sess).

@fchollet any keras references for this TF distributed computing support? Can we setup keras cluster as TF like this below?

model.ClusterSpec({"local": ["10.0.0.1", "10.0.0.2"]})
model.fit(x, y)

An example on how to user K.set_session(sess) for this would be awesome. In particular, in the examples folder.

This issue has been automatically marked as stale because it has not had recent activity. It will be closed after 30 days if no further activity occurs, but feel free to re-open a closed issue if needed.

Was this page helpful?
0 / 5 - 0 ratings