# BiOST

**Repository Path**: dream_chasing/BiOST

## Basic Information

- **Project Name**: BiOST
- **Description**: No description available
- **Primary Language**: Unknown
- **License**: Not specified
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2021-03-24
- **Last Updated**: 2021-03-24

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# Pytorch implementation of [Bidirectional One-Shot Unsupervised Domain Mapping (ICCV 2019)](https://arxiv.org/abs/1909.01595)

## Training

### Download Dataset
To download dataset : bash datasets/download_cyclegan_dataset.sh $DATASET_NAME where DATASET_NAME is one of (monet2photo, summer2winter_yosemite)


### Phase I - train VAE for domain B

python train.py --dataroot=./datasets/summer2winter_yosemite/trainB --name=summer2winter_yosemite_autoencoder --model=autoencoder --dataset_mode=single


For reverse direction:

 python train.py --dataroot=./datasets/summer2winter_yosemite/trainA --name=summer2winter_yosemite_autoencoder_reverse --model=autoencoder --dataset_mode=single


### Phase II - Train domain A and domain B together

python train.py --dataroot=./datasets/summer2winter_yosemite/ --name=summer2winter_yosemite_biost --load_dir=summer2winter_yosemite_autoencoder --model=biost --start=0

For reverse direction:

python train.py --dataroot=./datasets/summer2winter_yosemite/ --name=summer2winter_yosemite_biost --load_dir=summer2winter_yosemite_autoencoder --model=biost --A='B' --B='A --start=0