Displaying 1 to 5 from 5 results

T2F - T2F: text to face generation using Deep Learning

  •    Python

Text-to-Face generation using Deep Learning. This project combines two of the recent architectures StackGAN and ProGAN for synthesizing faces from textual descriptions. The project uses Face2Text dataset which contains 400 facial images and textual captions for each of them. The data can be obtained by contacting either the RIVAL group or the authors of the aforementioned paper. The code is present in the implementation/ subdirectory. The implementation is done using the PyTorch framework. So, for running this code, please install PyTorch version 0.4.0 before continuing.

MSG-GAN - MSG-GAN: Multi-Scale Gradients GAN (Architecture inspired from ProGAN but doesn't use layer-wise growing)

  •    Python

MSG-GAN (Multi-Scale Gradients GAN): A Network architecture inspired from the ProGAN. The architecture of this gan contains connections between the intermediate layers of the singular Generator and the Discriminator. The network is not trained by progressively growing the layers. All the layers get trained at the same time.

pro_gan_pytorch - ProGAN package implemented as an extension of PyTorch nn.Module

  •    Python

1.) Install your appropriate version of PyTorch. The torch dependency in this package uses the most basic "cpu" version. follow instructions on http://pytorch.org to install the "gpu" version of PyTorch. Use the modules pg.Generator, pg.Discriminator and pg.ProGAN. Mostly, you'll only need the ProGAN module.




pro_gan_pytorch-examples - Examples trained using the python pytorch package pro-gan-pth

  •    Python

There are two examples presented here for LFW dataset and MNIST dataset. Please refer to the following sections for how to train and / or load the provided trained weights for these models. Before running any of the following training experiments, please setup your VirtualEnv with the required packages for this project. Importantly, please install the progan package using $ pip install pro-gan-pth and your appropriate gpu / cpu version of PyTorch 0.4.0. Once this is done, you can proceed with the following experiments.