Hi everyone!
After months of effort to improve the sustainability of TensorFlow Java with a more robust CI/CD and better compatibility with TensorFlow native libraries, we are pleased to announce that TensorFlow Java version 1.0.0 has now been released and works better than ever!
TF Java now supports the same set of platforms as TensorFlow, including Linux and Mac arm64 systems. As this is the first major release of this library, we will conform to TensorFlow’s semantic version stability guarantees for future releases.
TF Java 1.0.0 is based on TensorFlow 2.16.2, and soon we’ll release TF-Java 1.1.0 which incorporates the latest TensorFlow release (2.18.0).
Please stay tuned for more updates and exciting new features to come in future versions TensorFlow Java, thank you for using it in your Java/Scala/Kotlin applications!
– SIG JVM
Note that due to TF’s deprecation of macOS x86_64 builds the TF-Java 1.0.0 release will be the last release for that platform.