mirror of
https://github.com/helm/charts.git
synced 2026-08-23 06:17:18 +00:00
Szymon Piotr KrasuskiandKubernetes Prow Robot
19b4b54a39
* Add custom variable to set number of trainign steps. By default script run in cheyang/distributed-tensorflow container runs with flag train_steps set to 0 which means that script will be running inifitely, However, it's to possible to set this flag to fixed number which will make script to run only passed number of training steps and stop. This change adds custom variable in hyperparams that lets user set custom value for training steps. Signed-off-by: Szymon <krasuski.szymon.piotr@gmail.com> * Bump chart version Signed-off-by: Szymon <krasuski.szymon.piotr@gmail.com>
13 lines
339 B
YAML
13 lines
339 B
YAML
apiVersion: v1
|
|
description: A Helm chart for running distributed TensorFlow on Kubernetes
|
|
name: distributed-tensorflow
|
|
version: 0.1.2
|
|
appVersion: 1.6.0
|
|
sources:
|
|
- https://github.com/tensorflow/tensorflow
|
|
- https://www.tensorflow.org/deploy/distributed
|
|
home: https://www.tensorflow.org
|
|
maintainers:
|
|
- name: cheyang
|
|
email: cheyang@163.com
|