TensorFlow Serving was developed and open sourced by the Google Brain team in 2015. It uses a standard architecture and set of APIs on which you can deploy machine learning algorithms, and is compatible with both TensorFlow models such as Inception-v3 and other types of models and data.
The Bitnami TensorFlow Serving stack comes with the Inception-v3 framework pre-installed and configured. Inception-v3 was developed for classifying complete images into 1,000 classes (such as llama, zebra, aircraft carrier, or even Pembroke Welsh Corgi) as part of the ImageNet Large Visual Recognition Challenge. You can use TensorFlow Serving for tasks like captioning images out-of-the-box, or you can add/create your own models and serve them instead.
We have a number of get started guides for using TensorFlow Serving in the cloud, on your Linux laptop, or even on your Kubernetes cluster! It has never been easier to start experimenting with this exciting technology, and the software is absolutely free to use. Give it a try!