How deepjavalibrary/djl-spring-boot-starter works
DJL Spring Boot Starter lets Spring applications configure DJL inference through dependency choices and djl.* properties. The documented flow is: select an engine starter, configure model criteria, let Spring auto-configuration load a model, then obtain a predictor for inference. The source excerpt confirms property-to-model configuration and predictor creation; sample application internals and engine starter POM contents were not provided, so those connections are limited to README-supported statements.
Main entry points: DjlAutoConfiguration.java Model autoconfig · DjlConfigurationProperties.java DJL properties · djl-spring-boot-starter-mxnet-auto/pom.xml MXNet auto · djl-spring-boot-starter-mxnet-linux-x86_64/pom.xml MXNet Linux
Components 15
Engine starters
- MXNet auto
djl-spring-boot-starter-mxnet-auto/pom.xml - MXNet Linux
djl-spring-boot-starter-mxnet-linux-x86_64/pom.xml - MXNet macOS
djl-spring-boot-starter-mxnet-osx-x86_64/pom.xml - PyTorch auto
djl-spring-boot-starter-pytorch-auto/pom.xml - TensorFlow auto
djl-spring-boot-starter-tensorflow-auto/pom.xml
Spring configuration
- DJL properties
djl-spring-boot-starter-autoconfigure/src/main/java/ai/djl/spring/configuration/DjlConfigurationProperties.java - Application types
djl-spring-boot-starter-autoconfigure/src/main/java/ai/djl/spring/configuration/ApplicationType.java - Model autoconfig
djl-spring-boot-starter-autoconfigure/src/main/java/ai/djl/spring/configuration/DjlAutoConfiguration.java
Inference runtime
- DJL criteria
- Model Zoo
- Loaded model
djl-spring-boot-starter-autoconfigure/src/main/java/ai/djl/spring/configuration/DjlAutoConfiguration.java - Predictor provider
djl-spring-boot-starter-autoconfigure/src/main/java/ai/djl/spring/configuration/DjlAutoConfiguration.java - Inference result
Other components
- Application developer
- Spring application
Connections 16
- Application developer → Spring application: configures
README.md - Application developer → MXNet auto: selects engine
README.md - Application developer → MXNet Linux: selects engine
README.md - Application developer → MXNet macOS: selects engine
README.md - Application developer → PyTorch auto: selects engine
README.md - Application developer → TensorFlow auto: selects engine
README.md - Spring application → DJL properties: sets properties
README.md - Model autoconfig → DJL properties: binds properties
djl-spring-boot-starter-autoconfigure/src/main/java/ai/djl/spring/configuration/DjlAutoConfiguration.java - Model autoconfig → Application types: reads application type
djl-spring-boot-starter-autoconfigure/src/main/java/ai/djl/spring/configuration/DjlAutoConfiguration.java - Model autoconfig → DJL criteria: builds criteria
djl-spring-boot-starter-autoconfigure/src/main/java/ai/djl/spring/configuration/DjlAutoConfiguration.java - Model autoconfig → Loaded model: loads model
djl-spring-boot-starter-autoconfigure/src/main/java/ai/djl/spring/configuration/DjlAutoConfiguration.java - DJL criteria → Model Zoo: selects model
- Model autoconfig → Model Zoo: lists models
djl-spring-boot-starter-autoconfigure/src/main/java/ai/djl/spring/configuration/DjlAutoConfiguration.java - Predictor provider → Loaded model: creates predictor
djl-spring-boot-starter-autoconfigure/src/main/java/ai/djl/spring/configuration/DjlAutoConfiguration.java - Spring application → Predictor provider: requests inference
- Predictor provider → Inference result: returns prediction