Auto-Parallel Technologies, Inc.

Perl TensorFlow is a programming interface which allows Perl developers to utilize Google's TensorFlow machine learning platform natively in their own Perl source code.

Features of the Perl TensorFlow interface include:

• 100% Compatible with Google TensorFlow
• Load & Run Neural Network Models
• Classify Unknown Objects
• Predict Future Values
• Native Perl API

Perl TensorFlow provides an easy-to-follow Perl TensorFlow Tutorial with real Perl source code examples and a runnable demo of image classification. The tutorial demonstrates a pre-built neural network model, utilizing the MobileNet V2 architecture and trained on the Imagenet dataset.

Features of the Perl TensorFlow Tutorial include:

• Available as both POD & Jupyter Notebook
• Real runnable Perl source code
• Download neural model from Google
• Download test images from Wikipedia
• User-friendly graphical output mode

The Perl TensorFlow Tutorial produces output showing the top 5 potential classifications of each input test image. Where possible, specific plant & animal breeds are automatically determined.

Each potential image class is given a statistical probability ranging from 0 (impossible) to 1 (certain).
All probabilities for a given image should add up to a total value of 1.0, representing 100% certainty overall.

Machine learning image classification can be used for a wide variety of academic & commercial applications, such as machine vision, warehouse inventory control, biomedical research, facial recognition, and wildlife conservation, to name just a few.

Features of the Perl TensorFlow Image Classification includes:

• Impressive visual demonstration of machine learning
• Clickable links to the original Wikipedia input files
• Statistical probabilities for each possible classification
• Variety of pre-defined sample images
• Easy to read graphical formatting