examples/titanic-kaggle-competition/titanic-kaggle-competition-...

162 lines
7.8 KiB
YAML

apiVersion: argoproj.io/v1alpha1
kind: Workflow
metadata:
generateName: boston-housing-pipeline-
annotations: {pipelines.kubeflow.org/kfp_sdk_version: 1.1.2, pipelines.kubeflow.org/pipeline_compilation_time: '2022-06-06T15:35:20.931416',
pipelines.kubeflow.org/pipeline_spec: '{"description": "An example pipeline that
trains and logs a regression model.", "name": "Boston Housing Pipeline"}'}
labels: {pipelines.kubeflow.org/kfp_sdk_version: 1.1.2}
spec:
entrypoint: boston-housing-pipeline
templates:
- name: bayes
container:
args: [--train_pickle, /tmp/inputs/input-0/data, --train_label, /tmp/inputs/input-1/data]
image: hubdocker76/titanic-bayes:v6
inputs:
artifacts:
- {name: featureengineering-train_pickle_out, path: /tmp/inputs/input-0/data}
- {name: featureengineering-train_label_out, path: /tmp/inputs/input-1/data}
outputs:
artifacts:
- {name: bayes-bayes_acc, path: /app/bayes_acc.txt}
- name: boston-housing-pipeline
dag:
tasks:
- name: bayes
template: bayes
dependencies: [featureengineering]
arguments:
artifacts:
- {name: featureengineering-train_label_out, from: '{{tasks.featureengineering.outputs.artifacts.featureengineering-train_label_out}}'}
- {name: featureengineering-train_pickle_out, from: '{{tasks.featureengineering.outputs.artifacts.featureengineering-train_pickle_out}}'}
- name: decision-tree
template: decision-tree
dependencies: [featureengineering]
arguments:
artifacts:
- {name: featureengineering-train_label_out, from: '{{tasks.featureengineering.outputs.artifacts.featureengineering-train_label_out}}'}
- {name: featureengineering-train_pickle_out, from: '{{tasks.featureengineering.outputs.artifacts.featureengineering-train_pickle_out}}'}
- name: featureengineering
template: featureengineering
dependencies: [preprocess-data]
arguments:
artifacts:
- {name: preprocess-data-test_pickle, from: '{{tasks.preprocess-data.outputs.artifacts.preprocess-data-test_pickle}}'}
- {name: preprocess-data-train_pickle, from: '{{tasks.preprocess-data.outputs.artifacts.preprocess-data-train_pickle}}'}
- {name: preprocess-data, template: preprocess-data}
- name: random-forest
template: random-forest
dependencies: [featureengineering]
arguments:
artifacts:
- {name: featureengineering-train_label_out, from: '{{tasks.featureengineering.outputs.artifacts.featureengineering-train_label_out}}'}
- {name: featureengineering-train_pickle_out, from: '{{tasks.featureengineering.outputs.artifacts.featureengineering-train_pickle_out}}'}
- name: regression
template: regression
dependencies: [featureengineering]
arguments:
artifacts:
- {name: featureengineering-train_label_out, from: '{{tasks.featureengineering.outputs.artifacts.featureengineering-train_label_out}}'}
- {name: featureengineering-train_pickle_out, from: '{{tasks.featureengineering.outputs.artifacts.featureengineering-train_pickle_out}}'}
- name: results
template: results
dependencies: [bayes, decision-tree, random-forest, regression, svm]
arguments:
artifacts:
- {name: bayes-bayes_acc, from: '{{tasks.bayes.outputs.artifacts.bayes-bayes_acc}}'}
- {name: decision-tree-decision_tree_acc, from: '{{tasks.decision-tree.outputs.artifacts.decision-tree-decision_tree_acc}}'}
- {name: random-forest-random_forest_acc, from: '{{tasks.random-forest.outputs.artifacts.random-forest-random_forest_acc}}'}
- {name: regression-regression_acc, from: '{{tasks.regression.outputs.artifacts.regression-regression_acc}}'}
- {name: svm-svm_acc, from: '{{tasks.svm.outputs.artifacts.svm-svm_acc}}'}
- name: svm
template: svm
dependencies: [featureengineering]
arguments:
artifacts:
- {name: featureengineering-train_label_out, from: '{{tasks.featureengineering.outputs.artifacts.featureengineering-train_label_out}}'}
- {name: featureengineering-train_pickle_out, from: '{{tasks.featureengineering.outputs.artifacts.featureengineering-train_pickle_out}}'}
- name: decision-tree
container:
args: [--train_pickle, /tmp/inputs/input-0/data, --train_label, /tmp/inputs/input-1/data]
image: hubdocker76/titanic-decision-tree:v1
inputs:
artifacts:
- {name: featureengineering-train_pickle_out, path: /tmp/inputs/input-0/data}
- {name: featureengineering-train_label_out, path: /tmp/inputs/input-1/data}
outputs:
artifacts:
- {name: decision-tree-decision_tree_acc, path: /app/decision_tree_acc.txt}
- name: featureengineering
container:
args: [--train_pickle, /tmp/inputs/input-0/data, --test_pickle, /tmp/inputs/input-1/data]
image: hubdocker76/titanic-feature-engineering:v8
inputs:
artifacts:
- {name: preprocess-data-train_pickle, path: /tmp/inputs/input-0/data}
- {name: preprocess-data-test_pickle, path: /tmp/inputs/input-1/data}
outputs:
artifacts:
- {name: featureengineering-train_label_out, path: /app/train_label_v2}
- {name: featureengineering-train_pickle_out, path: /app/train_v2}
- name: preprocess-data
container:
args: []
image: hubdocker76/titanic-pre-process-data:v9
outputs:
artifacts:
- {name: preprocess-data-test_pickle, path: /app/test}
- {name: preprocess-data-train_pickle, path: /app/train}
metadata:
labels:
kaggle-secret: "true"
- name: random-forest
container:
args: [--train_pickle, /tmp/inputs/input-0/data, --train_label, /tmp/inputs/input-1/data]
image: hubdocker76/titanic-randomforest:v4
inputs:
artifacts:
- {name: featureengineering-train_pickle_out, path: /tmp/inputs/input-0/data}
- {name: featureengineering-train_label_out, path: /tmp/inputs/input-1/data}
outputs:
artifacts:
- {name: random-forest-random_forest_acc, path: /app/random_forest_acc.txt}
- name: regression
container:
args: [--train_pickle, /tmp/inputs/input-0/data, --train_label, /tmp/inputs/input-1/data]
image: hubdocker76/titanic-logistic-regression:v5
inputs:
artifacts:
- {name: featureengineering-train_pickle_out, path: /tmp/inputs/input-0/data}
- {name: featureengineering-train_label_out, path: /tmp/inputs/input-1/data}
outputs:
artifacts:
- {name: regression-regression_acc, path: /app/regression_acc.txt}
- name: results
container:
args: [--bayes_acc, /tmp/inputs/input-0/data, --regression_acc, /tmp/inputs/input-1/data,
--random_forest_acc, /tmp/inputs/input-2/data, --decision_tree_acc, /tmp/inputs/input-3/data,
--svm_acc, /tmp/inputs/input-4/data]
image: hubdocker76/titanic-results:v9
inputs:
artifacts:
- {name: bayes-bayes_acc, path: /tmp/inputs/input-0/data}
- {name: regression-regression_acc, path: /tmp/inputs/input-1/data}
- {name: random-forest-random_forest_acc, path: /tmp/inputs/input-2/data}
- {name: decision-tree-decision_tree_acc, path: /tmp/inputs/input-3/data}
- {name: svm-svm_acc, path: /tmp/inputs/input-4/data}
- name: svm
container:
args: [--train_pickle, /tmp/inputs/input-0/data, --train_label, /tmp/inputs/input-1/data]
image: hubdocker76/titanic-svm:v2
inputs:
artifacts:
- {name: featureengineering-train_pickle_out, path: /tmp/inputs/input-0/data}
- {name: featureengineering-train_label_out, path: /tmp/inputs/input-1/data}
outputs:
artifacts:
- {name: svm-svm_acc, path: /app/svm_acc.txt}
arguments:
parameters: []
serviceAccountName: pipeline-runner