🏠 On-Premise#

The On-Premise installation of Giskard is often adapted if your data and model are private and/or you don’t have the possibility to use the cloud (for instance, because of connectivity issues). For other ways to install the Giskard Hub, check the Hugging Face Space setup (easy installation) or the Private Cloud installation.

The Giskard Hub is the app adapted for an enterprise use of Giskard. Extending the features of the open-source library, it enables you to:

  • Debug tests to diagnose your issues

  • Create domain-specific tests thanks to automatic model insights

  • Compare models to decide which model to promote

  • Collect business feedback of your model results

  • Share your results with your colleagues for alignment

  • Store all your QA objects (tests, data slices, evaluation criteria, etc.) in one place to work more efficiently

To know more about the 3 different licenses of the Hub (Trial, Startup and Enterprise) and its differences between the open-source library, have a look at this page.

1. Start the Hub#

To run the Giskard hub you need 3 requirements:

  1. A Linux, macOS machine, or WSL2 in Windows

  2. To install the Giskard Python library, see here.

  3. A running docker. After installation of Docker, you can run it in the background by just opening the Docker app (Mac or Windows).

For an easy installation of Docker you can execute:

sudo curl -fsSL https://get.docker.com -o get-docker.sh
sudo sh get-docker.sh

If you don’t have the sudo rights to run docker, please see the Docker setup page.

To start the Giskard hub, once the 3 above requirements are met, execute the following command in your terminal:

giskard hub start

You’ll then be able to open Giskard at http://localhost:19000/

⚠️ Warning

  • Make sure to run Docker before starting the Giskard hub

  • If the giskard command is not found then you need first to install the Giskard Python library (see the doc section here).

  • To see the available commands of the giskard hub, you can execute:

giskard hub --help

2. Start the ML worker#

Giskard executes your model using a worker that runs the model directly in your Python environment, with all the dependencies required by your model. You can either execute the ML worker:

  • From your local notebook within the kernel that contains all the dependencies of your model

  • From Google Colab within the kernel that contains all the dependencies of your model

  • Or from your terminal within the Python environment that contains all the dependencies of your model

Note

If you plan to use LLM-assisted tests or transformations, don’t forget to set the OPENAI_API_KEY environment variable before starting the Giskard worker.

To start the ML worker from your notebook, run in the cell of your notebook:

!giskard worker start -d -k YOUR_KEY

The API Access Key (YOUR_KEY) can be found in the Settings tab of the Giskard Hub.

⚠️ Warning

To see the available commands of the worker, you can execute:

!giskard worker --help

You’re all set to try Giskard in action. Upload your first model, dataset or test suite by following the upload an object page.

To start the ML worker from your Colab notebook, read the following instructions in order to get the ngrok_API_token. Once you got your token, run in your local terminal (not the the terminal from Colab):

giskard hub expose --ngrok-token <ngrok_API_token>

Then run in a cell of your Colab notebook:

!giskard worker start -d -k YOUR_KEY -u https://e840-93-23-184-184.ngrok-free.app

The API Access Key (YOUR_KEY) can be found in the Settings tab of the Giskard Hub.

⚠️ Warning

To see the available commands of the worker, you can execute:

!giskard worker --help

You’re all set to try Giskard in action. Upload your first model, dataset or test suite by following the upload an object page.

Run this command within the Python environment that contains all the dependencies of your model:

giskard worker start -k -u http://localhost:19000/

You then will be asked to provide your API Access Key. The API Access key can be found in the Settings tab of the Giskard hub (accessible via: http://localhost:19000/)

⚠️ Warning

To see the available commands of the worker, you can execute:

!giskard worker --help

You’re all set to try Giskard in action. Upload your first model, dataset or test suite by following the upload an object page.