Automate deployment, simplify code for provisioning to achieve speed and accuracy.
Original training file "cloud-ml-housing-prices.ipynb" was simplified (removed markdown and extracted bash commands into bash shell script) and only left with Jupyter notebook executable syntax.
A single bash script was created with extracted bash commands from original training file and were coded systematically to be executed as per lab requirement.
- Visit lab URL (https://google.qwiklabs.com/focuses/3644?parent=catalog) and login to your Qwiklabs account
- Launch this lab
- Open Google Console
- Accept Google Cloud Platform Terms of Service
- Activate Cloud Shell
- On Cloud Shell prompt, type following command and hit enter
source <(curl -s https://raw.githubusercontent.com/donchai/Predict-Housing-Prices-with-Tensorflow-and-AI-Platform/master/predict.sh) - Follow on-screen instructions on Cloud Shell to complete this lab
- Accept Google Cloud Platform Terms of Service
- Stage 1 "Create Storage Bucket"
- Stage 2 "Create the AI Platform notebook instance"
- Stage 3 "Download lab notebook"
- Stage 4 Before Complete "Train and deploy the Model for Predictions"
- Stage 4 After Complete "Train and deploy the Model for Predictions"
Contributors are welcome to participate in this project to enhance the deployment scripts in order to achieve full automation e.g.
-- Interact with Google Console via command line (to obtain username and password for automation after launch lab instead of manually copy and paste)
-- Accepting Google Cloud Platform Terms of Service via command line
-- Optimise the scripts to speed up the deployment time
-- Others
Unable to accept Google Cloud Platform Terms of Service from command line, still finding/exploring alternative to automate this manual step.
Login to Google Console and accept Google Cloud Platform Terms of Service.
https://google.qwiklabs.com/focuses/3644?parent=catalog
- File extension with .orig refer to original file without modification/simplification for reference purpose, not in use during deployment.
- File extension with .bak refer to development file for future enhancement reference - not in use during deployment.
Senpai Gil - The Linux Guru for sharing command tip and trick.