Input Optimizer is a PyTorch code designed to optimize input images for any deep learning network based on a specific target. The code was developed in the Human and Robot Interaction Laboratory (Taarlab).
One of the library's features is the ability to optimize input noise going through a pre-trained network on Mark1501 dataset. The following images demonstrate how the library optimizes based on a specific attribute:
|
|
Another feature of Input Optimizer is the ability to optimize input noise going through a pre-trained AlexNet model. The following images demonstrate how the library optimizes based on a specific attribute:
To use Input Optimizer, you need to have the following Python packages installed:
pythonCopy codetorch torchvision copy numpy PIL
For more information, please contact the developer at farbodsiahkali80@ut.ac.ir or farbodsiahkali80@gmail.com.