|
|
--- |
|
|
base_model: |
|
|
- microsoft/resnet-50 |
|
|
datasets: |
|
|
- wltjr1007/DomainNet |
|
|
language: |
|
|
- en |
|
|
license: mit |
|
|
metrics: |
|
|
- accuracy |
|
|
pipeline_tag: image-classification |
|
|
library_name: pytorch |
|
|
--- |
|
|
|
|
|
arxiv.org/abs/2505.24216 |
|
|
|
|
|
[Code](https://github.com/PrasannaPulakurthi/SPM) |
|
|
|
|
|
Paper: https://ieeexplore.ieee.org/document/11084606 |
|
|
|
|
|
# Shuffle PatchMix |
|
|
|
|
|
This is the official implementation of the **ICIP 2025** paper **"Shuffle PatchMix Augmentation with Confidence-Margin Weighted Pseudo-Labels for Enhanced Source-Free Domain Adaptation"**, by [Prasanna Reddy Pulakurthi](https://www.prasannapulakurthi.com/), [Majid Rabbani](https://www.rit.edu/directory/mxreee-majid-rabbani), [Jamison Heard](https://www.rit.edu/directory/jrheee-jamison-heard), [Sohail A. Dianat](https://www.rit.edu/directory/sadeee-sohail-dianat), [Celso M. de Melo](https://celsodemelo.net/), and [Raghuveer Rao](https://ieeexplore.ieee.org/author/37281258600). |
|
|
|
|
|
Shuffle PatchMix (SPM) is an augmentation technique that shuffles and blends image patches to generate diverse and challenging augmentations. It is combined with a novel reweighting strategy that prioritizes reliable pseudo-labels to mitigate label noise. |
|
|
|
|
|
## Installation |
|
|
|
|
|
1. Clone this repository. |
|
|
|
|
|
```bash |
|
|
git clone https://github.com/PrasannaPulakurthi/SPM.git |
|
|
cd SPM |
|
|
``` |
|
|
|
|
|
2. Install requirements using Python 3.9. |
|
|
|
|
|
```bash |
|
|
conda create -n spm-env python=3.9 |
|
|
conda activate spm-env |
|
|
``` |
|
|
|
|
|
3. The code is tested with [Pytorch](https://pytorch.org/get-started/locally/) 1.7.1, CUDA 11.0. |
|
|
|
|
|
```bash |
|
|
pip install torch==1.7.1+cu110 torchvision==0.8.2+cu110 -f https://download.pytorch.org/whl/torch_stable.html |
|
|
``` |
|
|
|
|
|
4. Please also make sure to install additional packages using the following command. |
|
|
|
|
|
```bash |
|
|
pip install -r requirements.txt |
|
|
``` |
|
|
|
|
|
## Datasets |
|
|
|
|
|
The model was evaluated on three major benchmarks: PACS, VisDA-C, and DomainNet-126. Instructions for preparing these datasets can be found in the Github repository. |
