Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Support GPU image for dnntrainer component #530

Merged
merged 8 commits into from
Dec 19, 2018

Conversation

hongye-sun
Copy link
Contributor

@hongye-sun hongye-sun commented Dec 13, 2018

Changes:

  • Parameterize dnntrainer build script to support build gpu image
  • Update dnntrainer dockerfile to base from tensorflow image
  • Update build script to push gpu image to gcr and update tags in release script

Future changes:

  • Update tfx taxi predict sample and notebook sample to support gpu mode

This change is Reviewable

@@ -42,12 +50,18 @@ if [ -z "${TAG_NAME}" ]; then
TAG_NAME=$(date +v%Y%m%d)-$(git describe --tags --always --dirty)-$(git diff | shasum -a256 | cut -c -6)
fi

TF_VERSION=1.6.0
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

TF images have version, CPU/GPU and python version. Maybe we should just add the base image tag command line parameter.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

done

@hongye-sun
Copy link
Contributor Author

/test kubeflow-pipeline-unit-test

@@ -169,3 +169,11 @@ tensorflow-tensorboard,https://raw.githubusercontent.com/tensorflow/tensorboard/
tensorflow-data-validation,https://raw.githubusercontent.com/tensorflow/data-validation/master/LICENSE,Apache 2.0
tensorflow-metadata,https://raw.githubusercontent.com/tensorflow/metadata/master/LICENSE,Apache 2.0
defusedxml,https://raw.githubusercontent.com/tiran/defusedxml/master/LICENSE,PSF
backports.functools-lru-cache,https://raw.githubusercontent.com/jaraco/backports.functools_lru_cache/master/LICENSE,MIT
cycler,https://raw.githubusercontent.com/matplotlib/cycler/master/LICENSE,MIT
h5py,https://raw.githubusercontent.com/h5py/h5py/07f3c362887563c1040b10d2cbd8cca49ceac1b9/licenses/license.txt
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

https://raw.githubusercontent.com/h5py/h5py/master/licenses/license.txt

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

done

@Ark-kun
Copy link
Contributor

Ark-kun commented Dec 14, 2018

BTW, what errors do you get when you try using the GPU image on a non-GPU cluster?
Normally TF supports both CPU and GPU.

@Ark-kun
Copy link
Contributor

Ark-kun commented Dec 14, 2018

/lgtm

@hongye-sun
Copy link
Contributor Author

I vaguely remember that It's related to CUDA driver not found error.

@hongye-sun
Copy link
Contributor Author

/assign qimingj

@qimingj
Copy link
Contributor

qimingj commented Dec 15, 2018

/approve

@k8s-ci-robot
Copy link
Contributor

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: qimingj

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

1 similar comment
@k8s-ci-robot
Copy link
Contributor

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: qimingj

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

@hongye-sun
Copy link
Contributor Author

/test kubeflow-pipeline-e2e-test

@hongye-sun
Copy link
Contributor Author

@qimingj could you give me another lgtm? thanks.

@qimingj
Copy link
Contributor

qimingj commented Dec 19, 2018

/lgtm

@k8s-ci-robot k8s-ci-robot merged commit f3de359 into kubeflow:master Dec 19, 2018
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

Successfully merging this pull request may close these issues.

4 participants