Home

Awesome

PixelPyramids: Exact Inference Models from Lossless Image Pyramids

<p align="center"> <img width="320" height="150" src="/assets/celeba_256.png" hspace="30"> <img width="320" height="150" src="/assets/celeba1024.png" hspace="30"> </p>

This repository is the PyTorch implementation of the paper:

PixelPyramids: Exact Inference Models from Lossless Image Pyramids (ICCV 2021)

Shweta Mahajan and Stefan Roth

Requirements

The following code is written in Python 3.6.10 and CUDA 9.0.

Requirements:

To install requirements:

conda config --add channels pytorch
conda config --add channels anaconda
conda config --add channels conda-forge
conda config --add channels conda-forge/label/cf202003
conda create -n <environment_name> --file requirements.txt
conda activate <environment_name>

Datasets

The datasets used in this project are:

Training

The important keyword arguments for training are,

Please follow the following instructions for training:

  1. Train a model on CelebA-HQ-256,
   	python main.py --params_file './params/celeba_256.json' 
  1. The model is evaluated after every epoch

Generation and Validation

Samples and test results in bits/dim can be obtained using main.py. Generated samples are stored in the ./samples folder. Download the checkpoints to the ckpts folder.

Memory requirements

The models were trained on four nvidia V100 GPU with 32 GB memory. The levels can be trained in parallel with a maximum of 24GB memory per level.

Results

Evaluation on different datasets

bits/dim
CelebA-HQ_2560.61
CelebA-HQ_10240.58
LSUN_bedroom_1280.88
LSUN_church_1281.07
LSUN_tower_1280.95
ImageNet_1283.40

Bibtex

@inproceedings{pixelpyramids21iccv,
  title     = {PixelPyramids: Exact Inference Models from Lossless Image Pyramids},
  author    = {Mahajan, Shweta and Roth, Stefan},
  booktitle = {Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV)},
  year = {2021}
}