Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/woozch/DSBN
Official Implementation of "Domain Specific Batch Normalization for Unsupervised Domain Adaptation (CVPR2019)"
https://github.com/woozch/DSBN
Last synced: about 1 month ago
JSON representation
Official Implementation of "Domain Specific Batch Normalization for Unsupervised Domain Adaptation (CVPR2019)"
- Host: GitHub
- URL: https://github.com/woozch/DSBN
- Owner: woozch
- Created: 2018-11-25T14:43:10.000Z (about 6 years ago)
- Default Branch: master
- Last Pushed: 2023-11-02T22:11:35.000Z (about 1 year ago)
- Last Synced: 2024-05-18T22:33:54.005Z (7 months ago)
- Language: Python
- Homepage:
- Size: 99.6 KB
- Stars: 146
- Watchers: 4
- Forks: 28
- Open Issues: 1
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
- awesome-normalization-techniques - [Python Reference
README
# Domain-Specific Batch Normalization for Unsupervised Domain Adaptation (DSBN)
Pytorch implementation of Domain-Specific Batch Normalization for Unsupervised Domain Adaptation (CVPR2019).
Woong-Gi Chang, Tackgeun You, Seonguk Seo, Suha Kwak, Bohyung Han
The first author name has changed from Woong-Gi Chang to Woojae Chang.
## Citation
If you want to cite our work, follow the link [arXiv](https://arxiv.org/abs/1906.03950).
## Installation
We recommand to create conda virtualenv nameded pytorch-py36
```bash
conda create -n pytorch-py36 python=3.6
source activate pytorch-py36
```
* Install [PyTorch 1.3](http://pytorch.org) with Python3.6, cuda10.1
* Install other dependencies
```bash
conda install numpy scipy matplotlib cython h5py
conda install -c menpo opencv
```
* For visualization using tensorboard
```bash
pip install tensorboardX
pip install tensorflow
```
* For color log print
```bash
pip install coloredlogs
```
## Dataset
data directory looks like below:
```text
data
├── Office
│ └── domain_adaptation_images
│ ├── amazon
│ ├── dslr
│ └── webcam
├── Office-home
│ └── OfficeHomeDataset_10072016
│ ├── Art
│ ├── Clipart
│ ├── Product
│ └── RealWorld
└── VisDA
├── test
├── train
└── validation
```
Datasets links to download.
#### VisDA-C dataset
* Refer this [site](http://ai.bu.edu/visda-2017/) and download the dataset.
#### OFFICE-31
* Download from [here](https://drive.google.com/open?id=0B4IapRTv9pJ1WGZVd1VDMmhwdlE)
#### OFFICE-HOME
* Download from [here](http://hemanthdv.org/OfficeHome-Dataset/).
Change domain name from "Real World" to "RealWorld" (Remove space between two words).
## Training Examples
### VISDA2017
### Stage1 Training (training existing UDA model with DSBN)
This is a example script for training [MSTN](http://proceedings.mlr.press/v80/xie18c/xie18c.pdf) on visda 2017 dataset for stage1.
Use resnet101dsbn for resnet101 with domain-specific batchnorm
```bash
# DSCN
python trainval_multi.py --model-name resnet101dsbn --exp-setting visda --sm-loss --adv-loss --source-datasets train --target-datasets validation --batch-size 40 --save-dir output/resnet101dsbn_visda_stage1 --print-console
# cf. batchnorm
python trainval_multi.py --model-name resnet101 --exp-setting visda --sm-loss --adv-loss --source-datasets train --target-datasets validation --batch-size 40 --save-dir output/resnet101_visda_stage1 --print-console
```
After training you can get stage1 model at save-dir.
### Stage2 Training (self-training a new model with the model trained on stage1)
For stage1, use finetune for single source unsupervised domain adaptation, and finetune_multi for multi source setting.
This is a example script for training MSTN on visda 2017 dataset for stage2.
```bash
# DSCN
python finetune_multi.py --model-name resnet101dsbn --exp-setting visda --source-dataset train --target-dataset validation --pseudo-target-loss default_ensemble --no-lambda --teacher-model-path output/resnet101dsbn_visda_stage1/best_resnet101dsbn+None+i0_train2validation.pth --learning-rate 5e-5 --batch-size 40 --save-dir output/resnet101dsbn_visda_stage2 --print-console
# cf. batchnorm
python finetune_multi.py --model-name resnet101 --exp-setting visda --source-dataset train --target-dataset validation --pseudo-target-loss default_ensemble --no-lambda --teacher-model-path output/resnet101_visda_stage1/best_resnet101+None+i0_train2validation.pth --learning-rate 5e-5 --batch-size 40 --save-dir output/resnet101_visda_stage2 --print-console
```
## Testing
```bash
python evlauate_multi.py --model-path [model-path] # for multi-source setting
```
File name should follow the format:
"best_{model_name}+{jitter}+{infeatures}_{source_dataset}2{target_dataset}.pth"
example:
best_resnet101dsbn+None+i0_train2validation.pth