Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/lsjsj92/kubeflow_example
kubeflow example
https://github.com/lsjsj92/kubeflow_example
deep-learning dockerfile kubeflow kubeflow-pipelines machine-learning machine-learning-pipelines ml mlops python python3
Last synced: about 14 hours ago
JSON representation
kubeflow example
- Host: GitHub
- URL: https://github.com/lsjsj92/kubeflow_example
- Owner: lsjsj92
- Created: 2020-04-05T07:01:39.000Z (almost 5 years ago)
- Default Branch: master
- Last Pushed: 2021-06-26T01:10:06.000Z (over 3 years ago)
- Last Synced: 2023-03-05T12:58:03.892Z (almost 2 years ago)
- Topics: deep-learning, dockerfile, kubeflow, kubeflow-pipelines, machine-learning, machine-learning-pipelines, ml, mlops, python, python3
- Language: Python
- Homepage:
- Size: 42 KB
- Stars: 17
- Watchers: 2
- Forks: 8
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# kubeflow_example
This repo is made for kubeflow example
# How to use it?
- I wrote an explain on my blog
- iris : https://lsjsj92.tistory.com/581
- titanic : https://lsjsj92.tistory.com/586
- metrics_evaluation_and_check_condtion : https://lsjsj92.tistory.com/589
1. iris
- Make kubeflow pipeline for simple example : iris data
- Doing load data -> training data
- If you want to start this example do this:
1. Make each dokcer image(preprocessing, training)
2. Make kubeflow pipeline(pipeline.py)
3. Start kubeflow
4. Upload pipeline and start experiment
2. titanic
- Make kubeflow pipline for simple AWS example with titanic data
- The process of this kubeflow pipeline :
In preprocessing
1. Get data from AWS s3
2. Preprocessing data
3. Upload to AWS s3 after preprocessing data
Until here, doing in preprocessing
In train_model
4. Get preprocessing data from s3
5. Training model
6. Upload model to s3
3. metrics_evaluation_and_check_condtion
- Make kubeflow evaluation and check condition
- Use iris data
- Evaluation
- Make metrics json
- Print metrics in kubeflow
- in 2_model_training/training_model.py
- Check condition
- Use dsl.condition
- in pipeline.py
- Result
- 