Overview
TensorFlow Java provides a robust API for integrating TensorFlow into JVM environments. It supports both CPU and GPU execution, operating in graph or eager mode. It is designed to leverage the widespread use of Java, Scala, and Kotlin in enterprises for large-scale machine learning adoption. The core component is the tensorflow-core-platform artifact, which includes the Java Core API and native dependencies for supported platforms. Extensions offer specialized support like Intel MKL-DNN and CUDA. It allows for building, saving, loading, and executing TensorFlow models. Example applications include image classification using pre-trained convolutional neural networks, showcasing graph construction, model loading, and graph execution.
Common tasks