Example of distribute training with TensorFlow.

[keisuke-umezawa]

Motivation

Currently, we do not have an example of distribute training with TensorFllow. It is useful for TensorFlow users.

https://www.tensorflow.org/guide/distributed_training

Description

Create an example of distribute training with TensorFllow.

[hvy]

Still seems to be an open issue.

Just a note that we cannot test GPU or TPU code for the time being in our CI environment. @keisuke-umezawa could you maybe elaborate a bit on what type of distributed examples you had in mind? I think that'll make this issue easer to tackle.

[github-actions[bot]]

This issue has not seen any recent activity.

[toshihikoyanase]

I'm not familiar with tensorflow-ecosystem, but this Keras distributed training tutorial provides more concrete code example than the tensorflow's document. In the tutorial, a simple convolutional neural network is trained with MNIST dataset in parallel.

So, how about creating the keras distributed example based on the tutorial, and then translate it using tensorflow APIs? @keisuke-umezawa @hvy

CC @sfujiwara Please feel free to add any comments on this topic.

[keisuke-umezawa]

So, how about creating the keras distributed example based on the tutorial, and then translate it using tensorflow APIs?

That's a better idea!

[github-actions[bot]]

This issue has not seen any recent activity.

[github-actions[bot]]

This issue was closed automatically because it had not seen any recent activity. If you want to discuss it, you can reopen it freely.