Neural Api Versions Save

The main topics in this release are:

• All fully connected layers now support OpenCL on the forward pass.
• Better documentation.
• A new Autoencoder example that shows an autoencoder built with hyperbolic tangents and trained with Tiny ImageNet 200.

  

• Some small bug fixes.

This release includes a super resolution example: https://github.com/joaopauloschuler/neural-api/tree/master/examples/SuperResolution

There is also a command line tool so everyone can now increase image resolution with no more than FPC (no external library are required) via command line:

#SuperResolution -i street.png -o street3.png
Loading input file: street.png
Input image size: 158x214x3
Creating Neural Network...
Resizing with tiles...
Neural network file found at ../../../examples/SuperResolution : super-resolution-7-64-sep.nn
Padding input image.
Resizing with tiles to: 288x416x3
Saving output file: street3.png

This release also fixes bug #25 .

There are plenty of image datasets where each folder represents a class of images. To make the pascal coding simpler, a new procedure CreateVolumesFromImagesFromFolder has been added.

New procedure CreateVolumesFromImagesFromFolder has parallel code so classes are loaded into memory in parallel. I consider this code tremendously fast and it outperforms code that I saw in other APIs. I should say thank you to FPC developers for coding FPImage and plenty of other super fast bits and pieces that allowed me to implement a fast CreateVolumesFromImagesFromFolder.

Source code examples:

• Simple Plant Leaf Disease Image Classifier
• Tiny ImageNet 200

Some fixes were applied to neuralfit. We are approaching v1.0!

Few fixes have been made.

CIFAR-100, MNIST and Fashion MNIST are currently in testing.

Although testing continues, this version is almost ready for 1.0!