https://github.com/JinjingZhu/PMTrans

https://github.com/JinjingZhu/PMTrans

Last synced: 7 days ago
JSON representation

• Host: GitHub
• URL: https://github.com/JinjingZhu/PMTrans
• Owner: JinjingZhu
• Created: 2023-05-26T15:10:58.000Z (over 1 year ago)
• Default Branch: main
• Last Pushed: 2023-06-06T07:33:20.000Z (over 1 year ago)
• Last Synced: 2024-08-02T15:26:04.838Z (3 months ago)
• Language: Python
• Size: 4.18 MB
• Stars: 45
• Watchers: 2
• Forks: 7
• Open Issues: 12
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# PMTrans

Patch-Mix Transformer for Unsupervised Domain Adaptation: A Game Perspective

### CVPR 2023 Highlight

### There are some typos in results on the DomainNet dataset. And we revise these typos in the [newest version](https://arxiv.org/abs/2303.13434). Please check it.

This is a rough version, I will continue to polish it.

### Pretrained Swin-B

- Download [swin_base_patch4_window7_224_22k.pth](https://github.com/SwinTransformer/storage/releases/download/v1.0.0/swin_base_patch4_window7_224_22k.pth) and put it into `pretrained_models`

### Install

- Create a conda virtual environment and activate it:

```bash

conda create -n swin python=3.7 -y

conda activate swin

```

- Install `CUDA==10.1` with `cudnn7` following

  the [official installation instructions](https://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html)

- Install `PyTorch==1.7.1` and `torchvision==0.8.2` with `CUDA==10.1`:

```bash

conda install pytorch==1.7.1 torchvision==0.8.2 cudatoolkit=10.1 -c pytorch

```

- Install `timm==0.3.2`:

```bash

pip install timm==0.3.2

pip install tensorboard 

```

- Install `Apex`:

```bash

git clone https://github.com/NVIDIA/apex

cd apex

pip install -v --disable-pip-version-check --no-cache-dir ./

https://github.com/NVIDIA/apex/issues/1227

```

- Install other requirements:

```bash

pip install opencv-python==4.4.0.46 termcolor==1.1.0 yacs==0.1.8

```

### Datasets:

- Download the `Office31, Office Home, VisDA and Domainnet` Make a file recording the path and label of image like txt files in `datasets/office_home/`

 ```bash

  $ tree data

  datasets

  ├── ofice_home

  │   ├── Art.txt

  │   ├── Clipart.txt

  │   ├── Product.txt

  │   ├── Real_World.txt

  └── ...

  ```   

### Training:

bash dist_train.sh