Home > Product Manuals > Model Deploy > Model Deploy

Model Deploy

FastScore Model Deploy is a containerized Jupyter notebook server with FastScore’s model deployment and Jupyter integration toolkit built in. It is built on top of the Jupyter data science Docker image. Model Deploy provides model creation and deployment tools for R and Python (2 & 3) notebooks, as well as for PFA (through Python 2).

Starting Model Deploy

Start Model Deploy with the following command:

docker run -it --rm -p 8888:8888 fastscore/model-deploy:latest

If other services in the FastScore fleet are also running on the same host, it may be advantageous to start Model Deploy with the --net="host" option, so that these services are accessible from localhost.

Model Deploy may also be started with any of the additional configuration options available to the Jupyter base Docker image, see the documentation for more details.

Once the container is created, it will be accessible from port 8888 (by default) on the host machine, using the token generated during the startup process.

Install to Existing Jupyter Notebook

Model Deploy is also avaiable via the fastscoredeploy library and can be installed using pip install fastscoredeploy from within the Jupyter Notebook terminal.

Model Deploy functionality

Model Deploy provides a number of features to make it easy to migrate a model into FastScore:

Example notebooks demonstrating this functionality are included with the Model Deploy container.