Skip to content
master
Go to file
Code

Latest commit

 

Git stats

Files

Permalink
Failed to load latest commit information.
Type
Name
Latest commit message
Commit time
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

README.md

PixelCNN & PixelRNN in TensorFlow

TensorFlow implementation of Pixel Recurrent Neural Networks. This implementation contains:

model

  1. PixelCNN
  • Masked Convolution (A, B)
  1. PixelRNN
  • Row LSTM (in progress)
  • Diagonal BiLSTM (skew, unskew)
  • Residual Connections
  • Multi-Scale PixelRNN (in progress)
  1. Datasets
  • MNIST
  • cifar10 (in progress)
  • ImageNet (in progress)

Requirements

Usage

First, install prerequisites with:

$ pip install tqdm gym[all]

To train a pixel_rnn model with mnist data (slow iteration, fast convergence):

$ python main.py --data=mnist --model=pixel_rnn

To train a pixel_cnn model with mnist data (fast iteration, slow convergence):

$ python main.py --data=mnist --model=pixel_cnn --hidden_dims=64 --recurrent_length=2 --out_hidden_dims=64

To generate images with trained model:

$ python main.py --data=mnist --model=pixel_rnn --is_train=False

Samples

Samples generated with pixel_cnn after 50 epochs.

generation_2016_08_01_16_40_28.jpg

Training details

Below results uses two different parameters

[1] --hidden_dims=16 --recurrent_length=7 --out_hidden_dims=32
[2] --hidden_dims=64 --recurrent_length=2 --out_hidden_dims=64

Training results of pixel_rnn with [1] (yellow) and [2] (green) with epoch as x-axis:

pixel_rnn

Training results of pixel_cnn with [1] (orange) and [2] (purple) with epoch as x-axis:

pixel_cnn

Training results of pixel_rnn (yellow, green) and pixel_cnn (orange, purple) with hour as x-axis:

pixel_rnn_cnn_relative

References

Author

Taehoon Kim / @carpedm20

Releases

No releases published

Languages

You can’t perform that action at this time.