weilinie

Bongard-LOGO

Bongard-LOGO is a Python code repository with the purpose of generating synthetic Bongard problems on a large scale with little human intervention.

DiffPure

A new adversarial purification method that uses the forward and reverse processes of diffusion models to remove adversarial perturbations.

LACE

A new play-and-plug method of controlling an existing generative model with conditioning attributes and their compositions.

BackpropVis

Part of the experiments in "A Theoretical Explanation for Perplexing Behaviors of Backpropagation-based Visualizations"

GAN-general

Tensorflow implementation for training GANs with various objectives and gradient penalties, different network architectures, both image and word generations

GAN-QA

Applying GANs in improving question generation and answering

GLO-tensorflow

tensorflow code for the paper "Optimizing the Latent Space of Generative Networks" https://arxiv.org/abs/1707.05776

JARE

Code for Towards a Better Understanding and Regularization of GAN Training Dynamics

RelGAN

Implementation of RelGAN: Relational Generative Adversarial Networks for Text Generation

Understand-GAN

weilinie/RelGAN

Implementation of RelGAN: Relational Generative Adversarial Networks for Text Generation

weilinie/GAN-QA

Applying GANs in improving question generation and answering

weilinie/BackpropVis

Part of the experiments in "A Theoretical Explanation for Perplexing Behaviors of Backpropagation-based Visualizations"

weilinie/GLO-tensorflow

tensorflow code for the paper "Optimizing the Latent Space of Generative Networks" https://arxiv.org/abs/1707.05776

weilinie/JARE

Code for Towards a Better Understanding and Regularization of GAN Training Dynamics

weilinie/GAN-general

Tensorflow implementation for training GANs with various objectives and gradient penalties, different network architectures, both image and word generations

weilinie/Understand-GAN

weilinie/grad_vis

weilinie/HoloGAN

HoloGAN

weilinie/website

weilinie/weilinie.github.io