DJL uses slf4j-api to print logs. DJL library itself does not define the logging framework. Instead, users have to choose their own logging framework at deployment time. Please refer to slf4j user manual to how to configure your logging framework.
If you didn’t add any logging framework to your project, you will see following out:
SLF4J: Failed to load class "org.slf4j.impl.StaticLoggerBinder".
SLF4J: Defaulting to no-operation (NOP) logger implementation
SLF4J: See http://www.slf4j.org/codes.html#StaticLoggerBinder for further details.
You can resolve this issue by adding a popular logging framework to your project.
For a quick prototyping, you can include slf4j-simple to your project to enable logging (slf4j-simple is not recommended for production deployment):
For Maven:
<dependency>
<groupId>org.slf4j</groupId>
<artifactId>slf4j-simple</artifactId>
<version>1.7.36</version>
</dependency>
For Gradle:
implementation "org.slf4j:slf4j-simple:1.7.36"
Then you can use system properties to configure slf4j-simple log level:
-Dorg.slf4j.simpleLogger.defaultLogLevel=debug
See SimpleLogger for more detail.
In our examples module, we use log4j2 binding. While using log4j2 binding, you also need add a log4j2.xml file.
If you want to use other logging framework such as logback
, you can just add the following dependency into your build.gradle:
implementation "ch.qos.logback:logback-classic:1.4.5"
or for Maven:
<dependency>
<groupId>ch.qos.logback</groupId>
<artifactId>logback-classic</artifactId>
<version>1.4.5</version>
<scope>test</scope>
</dependency>
log4j2
allows you to customize logging level using system properties. See our examples log4j2.xml.
With this configuration, you can easily change your logging level using java command line options:
cd examples
./gradlew run -Dai.djl.logging.level=debug
Change logging level to debug is particularly useful to troubleshooting following issues:
Debugging Model loading issue
When using ModelZoo API, sometime run into ModelNotFoundException
, and wondering
why the desired model doesn’t match the search criteria. DJL prints out model loading details at debug
level to help you nail down the issue.
Debugging Engine loading issue
DJL support multiple engines, engine initialization is highly depends on system environment. You might run into various problems. Using debug level logging can help you identify configuration issues.
Most of the logging issues are not DJL specific. Please refer to SLF4J official document.
log4j has 1.2 and 2.0 two major versions; and they are not compatible. Both of them might be indirectly included in your project and may cause class loading issue. Please make sure only one log4j version is included.
Many logging frameworks require a configuration file. Please make sure add your configuration file to the project.