https://github.com/mind-the-pineapple/autoregressive-models

Tensorflow 2.0 implementation of Deep Autoregressive Models
https://github.com/mind-the-pineapple/autoregressive-models

autoregressive-neural-networks gated-pixelcnn generative-models pixelcnn pixelrnn tensorflow tensorflow-models tensorflow2 vq-vae2

Last synced: 2 days ago
JSON representation

Tensorflow 2.0 implementation of Deep Autoregressive Models

• Host: GitHub
• URL: https://github.com/mind-the-pineapple/autoregressive-models
• Owner: Mind-the-Pineapple
• License: mit
• Created: 2019-06-25T23:00:36.000Z (over 5 years ago)
• Default Branch: master
• Last Pushed: 2021-09-14T12:57:06.000Z (over 3 years ago)
• Last Synced: 2025-01-31T03:01:32.243Z (12 days ago)
• Topics: autoregressive-neural-networks, gated-pixelcnn, generative-models, pixelcnn, pixelrnn, tensorflow, tensorflow-models, tensorflow2, vq-vae2
• Language: Jupyter Notebook
• Homepage:
• Size: 3.34 MB
• Stars: 28
• Watchers: 5
• Forks: 3
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
 Deep Autoregressive Models 







[![MIT license](http://img.shields.io/badge/license-MIT-brightgreen.svg)](https://github.com/Warvito/vq-vae/blob/master/LICENSE)

This repository is a supplement to our blog series *Deep Autoregressive Models*.

## Setup

Clone the git repository :

    git clone https://github.com/Mind-the-Pineapple/Autoregressive-models.git

Python 3 with [TensorFlow 2.0+](https://www.tensorflow.org/) are the primary requirements.

Install virtualenv and create a new virtual environment:

    sudo apt update

    sudo apt install python3-dev python3-pip

    sudo pip3 install -U virtualenv  # system-wide install

    virtualenv --system-site-packages -p python3 ./venv

Then, install requirements

    source ./venv/bin/activate

    pip3 install --upgrade pip

    pip3 install -r requirements.txt

--------------------------------------------------------------------------------------------------------------------------------------------------------------------

 1. Autoregressive Models — PixelCNN 



Creating digits with deep generative models!

- [PixelCNN Medium Story](https://towardsdatascience.com/autoregressive-models-pixelcnn-e30734ede0c1)

- [Google Colab](https://colab.research.google.com/github/Mind-the-Pineapple/Autoregressive-models/blob/master/1%20-%20Autoregressive%20Models%20-%20PixelCNN/pixelCNN.ipynb)

- [Code](https://github.com/Mind-the-Pineapple/Autoregressive-models/blob/master/1%20-%20Autoregressive%20Models%20-%20PixelCNN/pixelCNN.py)

- [Paper -> Pixel Recurrent Neural Networks](https://arxiv.org/abs/1601.06759)







 2. Modelling Coloured Images 



Extending the model to the full colour spectrum

- [PixelCNN - RGB Medium Story](https://blog.usejournal.com/modelling-coloured-images-acd0ebde0102)

- [Google Colab](https://colab.research.google.com/gist/PedroFerreiradaCosta/a770317efa23f36b3c5009a9f21169f3/pixelcnn-rgb.ipynb?authuser=1#scrollTo=deJgSHmOBGfk)

- [Code](https://github.com/Mind-the-Pineapple/Autoregressive-models/blob/master/2%20-%20Modelling%20data%20with%20multiple%20channels/pixelCNN_RGB.py)

- [Paper -> Pixel Recurrddent Neural Networks](https://arxiv.org/abs/1601.06759)










 3. PixelCNN's Blind Spot and how to fix it 



How to fix the Pixel CNN's blind spot

- [PixelCNN Medium Story](https://towardsdatascience.com/pixelcnns-blind-spot-84e19a3797b9)

- [Google Colab](https://colab.research.google.com/github/Mind-the-Pineapple/Autoregressive-models/blob/master/3%20-%20PixelCNNs%20Blind%20spot%20and%20Gated%20PixelCNNs/PixelCNN_Blind_spot.ipynb)

- [Code](https://github.com/Mind-the-Pineapple/Autoregressive-models/blob/master/3%20-%20PixelCNNs%20Blind%20spot%20and%20Gated%20PixelCNNs/pixelcnn_blind_spot.py)

- [Paper -> Gated PixelCNNs](https://arxiv.org/abs/1606.05328)