CI for deep learning

Prashant Bansod

Hi, I am working on Deep Learning project with my team of 3 people. I am pretty new to CI.  One of our team member( remote worker)  merged code to the development branch. The development branch apparently broke the build for the test environment used in the office. As we need to run our codes on GPU, I was wondering how can we build it using the `pipelines`. I would appreciate your suggestions. Pardon me for my technical knowledge. Thanks. 

Philip Hodder

Hello Prashant,

Currently, Bitbucket Pipelines does not have an option to run builds on a GPU. You can open a feature request to add support for that here: https://bitbucket.org/site/master/issues/new

The closest workaround I can think of would be to have your Pipelines trigger GPU-related tests to run on a cloud provider that provisions GPU-based instances. You can then read the result of that back into Pipelines and set the build result as appropriate, based on the final output.

Thanks,

Phil