stevedev7 / cloud-automl

GCP AutoML Demonstration

Create a service account {#service-account}

Create an account with AutoML Admin role and download the key as json file. After the credentials are downloaded set the environment variable GOOGLE_APPLICATION_CREDENTIALS as the absolute path to the credentials file.

$ GOOGLE_APPLICATION_CREDENTIALS="absolute/path/to/credentials"

Use this link to know how to set up one

Install dependencies

Run the following code in the terminal or virtual environment to install the required dependencies.

$ pip install -r requirements.txt --upgrade

Run the code

• Replace the project_id variable with your project_id on line 3 of index.py.

  project_id = "your_project_id"

• Replace gs bucket path to the dataset with your own path in the file index.py, line number 4.

  path = "gs://path/to/your/dataset"

• On line number 5 of index.py change the display_name variable to anything that you want it to be reflected on the dataset and model.

  display_name="<any_display_name>"

  NOTE Do not use any special character other than underscore ( _ )

• Set the appropriate target column on line number 6

  target_column_name = "Target Column"

• And then run the script using the following command.

  $ python index.py

  OR

  $ python3 index.py

Check the console for messeges

Go to Tables datasets to view your dataset.

Go to Tables models to view your model.