mirror of
https://github.com/hanzo-ml/katib.git
synced 2026-10-30 14:46:14 +00:00
Mahdi KhashanandGitHub
19db2d47d4
* move to ghcr Signed-off-by: mahdikhashan <[email protected]> * move images to ghcr Signed-off-by: mahdikhashan <[email protected]> * manifests Signed-off-by: mahdikhashan <[email protected]> * change registry in all path Signed-off-by: mahdikhashan <[email protected]> * update script Signed-off-by: mahdikhashan <[email protected]> * fix Signed-off-by: mahdikhashan <[email protected]> * fix Signed-off-by: mahdikhashan <[email protected]> * slight fix Signed-off-by: mahdikhashan <[email protected]> --------- Signed-off-by: mahdikhashan <[email protected]> Signed-off-by: Mahdi Khashan <[email protected]>
91 lines
3.0 KiB
YAML
91 lines
3.0 KiB
YAML
---
|
|
# This example shows how you can use trial metadata in template substitution
|
|
apiVersion: kubeflow.org/v1beta1
|
|
kind: Experiment
|
|
metadata:
|
|
namespace: kubeflow
|
|
name: trial-metadata-substitution
|
|
spec:
|
|
objective:
|
|
type: minimize
|
|
goal: 0.001
|
|
objectiveMetricName: loss
|
|
algorithm:
|
|
algorithmName: random
|
|
parallelTrialCount: 3
|
|
maxTrialCount: 12
|
|
maxFailedTrialCount: 3
|
|
parameters:
|
|
- name: lr
|
|
parameterType: double
|
|
feasibleSpace:
|
|
min: "0.01"
|
|
max: "0.05"
|
|
- name: momentum
|
|
parameterType: double
|
|
feasibleSpace:
|
|
min: "0.5"
|
|
max: "0.9"
|
|
trialTemplate:
|
|
primaryContainerName: training-container
|
|
trialParameters:
|
|
- name: learningRate
|
|
description: Learning rate for the training model
|
|
reference: lr
|
|
- name: momentum
|
|
description: Momentum for the training model
|
|
reference: momentum
|
|
- name: trialName
|
|
description: Name of the current trial's job
|
|
reference: ${trialSpec.Name}
|
|
- name: trialNamespace
|
|
description: Namespace of the current trial's job
|
|
reference: ${trialSpec.Namespace}
|
|
- name: trialKind
|
|
description: Kind of the current trial's job
|
|
reference: ${trialSpec.Kind}
|
|
- name: trialAPIVersion
|
|
description: API version of the current trial's job
|
|
reference: ${trialSpec.APIVersion}
|
|
- name: trialLabelCustom
|
|
description: Trial's job label with custom value
|
|
reference: ${trialSpec.Labels[custom-key]}
|
|
- name: trialAnnotationCustom
|
|
description: Trial's job annotation with custom value
|
|
reference: ${trialSpec.Annotations[custom-key]}
|
|
trialSpec:
|
|
apiVersion: batch/v1
|
|
kind: Job
|
|
metadata:
|
|
annotations:
|
|
"custom-key": "custom-annotation"
|
|
labels:
|
|
"custom-key": "custom-label"
|
|
spec:
|
|
template:
|
|
spec:
|
|
containers:
|
|
- name: training-container
|
|
image: ghcr.io/kubeflow/katib/pytorch-mnist-cpu:latest
|
|
command:
|
|
- "python3"
|
|
- "/opt/pytorch-mnist/mnist.py"
|
|
- "--epochs=1"
|
|
- "--batch-size=16"
|
|
- "--lr=${trialParameters.learningRate}"
|
|
- "--momentum=${trialParameters.momentum}"
|
|
env:
|
|
- name: TRIAL_NAME
|
|
value: ${trialParameters.trialName}
|
|
- name: TRIAL_NAMESPACE
|
|
value: ${trialParameters.trialNamespace}
|
|
- name: TRIAL_KIND
|
|
value: ${trialParameters.trialKind}
|
|
- name: TRIAL_API_VERSION
|
|
value: ${trialParameters.trialAPIVersion}
|
|
- name: TRIAL_LABEL_CUSTOM
|
|
value: ${trialParameters.trialLabelCustom}
|
|
- name: TRIAL_ANNOTATION_CUSTOM
|
|
value: ${trialParameters.trialAnnotationCustom}
|
|
restartPolicy: Never
|