Contact your DataRobot representative for information on enabling the Scoring Code feature.
The Scoring Code feature exports Scoring Code for qualifying Leaderboard models, allowing you to use DataRobot-generated models outside of the platform. From the Download tab, you can download both a pre-compiled JAR file (with all dependencies included) as well as the source code JAR file. All Scoring Code JARs contain prediction calculation logic identical to the DataRobot API—the code generation mechanism tests each model for accuracy as a part of the generation process.
Scoring Code that can be used anywhere where Java code can be executed.
Low-latency scoring without the API call overhead. Java code is typically faster than scoring through the Python API.
The ability to integrate models into systems that can’t necessarily communicate with the DataRobot API. It can be used either as a primary means of scoring for fully offline systems or as a backend for systems that are using the DataRobot API.
A complete match of predictions generated by DataRobot and the JAR model (whereas DataRobot Prime creates a model approximation).
The ability to use additional hardware to score large amounts of data.
Scoring Code verification¶
When the feature is enabled, DataRobot generates a Java alternative for each blueprint preprocessing step and compares its results on the validation set with the original results. If the difference between results is greater than 0.00001, DataRobot does not provide you an option to download the Scoring Code. In this way DataRobot ensures that the Scoring Code JAR model always produces the same predictions as the original model. If verification fails, check the Log tab for error details.
See also information on downloading Scoring Code.
See the following sections for more details:
- Scoring Code Quickstart
- Using the command line
- Downloading Scoring Code for existing projects
- Backward compatibility with the Java API
- The JAR file structure
- Using Scoring Code models in SageMaker
- Using Scoring Code models in Azure ML
The model JAR files require Java 8 or later.
Downloading Scoring Code¶
Models displaying the SCORING CODE icon on the Leaderboard are available for Scoring Code download.
The ability to download Scoring Code for a model from the Leaderboard depends on the pricing plan for your organization. Legacy users will see the option described below. MLOps users can download Scoring Code from the Leaderboard and directly from a deployment. Reference this documentation for more information.
Navigate to the Predict > Downloads tab, where you can select a download option and access a link to the up-to-date Java API documentation.
There are two download options for Scoring Code:
|Binary||These are compiled Java executables, which include all dependencies and can be used to make predictions.|
|Source (Java source code files)||These are a non-obfuscated version of the model; they cannot be used to score the model since they are not compiled and dependency packages are not included. Use the source files to explore the model’s decision making process.|
Additional information about the Java API can be found in the DataRobot javadocs.