Skip to content

Add support for multi-GPU training #1

Description

@TTitcombe

Tensorflow can allow distributed training strategies, such as multi-GPU on a single machine, or across multiple machines.

Furthermore, you can utilise Google TPUs if training on the cloud.

Information is here https://www.tensorflow.org/alpha/guide/distribute_strategy
Note not all strategies are currently supported in 2.0

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions