This repository contains PyTorch implementation of the Paper Training Triplet Networks With GAN (Triplet-GANs) on MNIST Dataset
- Batch Size: 100
- Pre-train learning rate: 0.0003
- Train learning rate: 0.0003
- Pre-train epochs: 100
- Training epochs: 30
- Input size of generator: 100
- Weight initialization to standard normal value with mean 0 and variance 0.05 for convolutional layers and to variance of 0.02 for fully connected layers.
- Weight normalization.
- Batch norm layers in initial layers of generator.
- Set output layer of generator to Sigmoid non-linearity.
- Use feature matching to calculate generator's loss.
Accuracy | |
---|---|
N=100, M=16 | 0.9806 |
N=100, M=32 | 0.9773 |
N=200, M=16 | 0.9817 |
TSNE-Plots After Training
- Improved techniques for training GANs. [Paper] (NeurIPS, 2016), [Code]
- Official Code Repo (Lasagne Code): https://github.com/maciejzieba/tripletGAN