[CVPR2021] Domain Consensus Clustering for Universal Domain Adaptation

Last update: Dec 26, 2022

Overview

[CVPR2021] Domain Consensus Clustering for Universal Domain Adaptation

Prerequisites

To install requirements:

pip install -r requirements.txt

Python 3.6
GPU Memory: 10GB
Pytorch 1.4.0

Getting Started

Download the dataset: Office-31, OfficeHome, VisDA, DomainNet.

Data Folder structure:

Your dataset DIR:
|-Office/domain_adaptation_images
| |-amazon
| |-webcam
| |-dslr
|-OfficeHome
| |-Art
| |-Product
| |-...
|-VisDA
| |-train
| |-validataion
|-DomainNet
| |-clipart
| |-painting
| |-...

You need you modify the data_path in config files, i.e., config.root

Training

Train on one transfer of Office:

CUDA_VISIBLE_DEVICES=0 python office_run.py note=EXP_NAME setting=uda/osda/pda source=amazon target=dslr

To train on six transfers of Office:

CUDA_VISIBLE_DEVICES=0 python office_run.py note=EXP_NAME setting=uda/osda/pda transfer_all=1

Train on OfficeHome:

CUDA_VISIBLE_DEVICES=0 python officehome_run.py note=EXP_NAME setting=uda/osda/pda source=Art target=Product

CUDA_VISIBLE_DEVICES=0 python officehome_run.py note=EXP_NAME setting=uda/osda/pda transfer_all=1

The final results (including the best and the last) will be saved in the ./snapshot/EXP_NAME/result.txt.

Notably, transfer_all wil consumes more shared memory.

Citation

If you find it helpful, please consider citing:

@inproceedings{li2021DCC,
  title={Domain Consensus Clustering for Universal Domain Adaptation},
  author={Li, Guangrui and Kang, Guoliang and Zhu, Yi and Wei, Yunchao and Yang, Yi},
  booktitle={IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
  year={2021}
}

[CVPR2021] Domain Consensus Clustering for Universal Domain Adaptation

Related tags

Overview

[CVPR2021] Domain Consensus Clustering for Universal Domain Adaptation

Prerequisites

Getting Started

Training

Citation

Owner

Guangrui Li

All public open-source implementations of convnets benchmarks

Neural Cellular Automata + CLIP

Proof-Of-Concept Piano-Drums Music AI Model/Implementation

A medical imaging framework for Pytorch

Official code for the paper: Deep Graph Matching under Quadratic Constraint (CVPR 2021)

Learning from graph data using Keras

Code release to accompany paper "Geometry-Aware Gradient Algorithms for Neural Architecture Search."

Efficient 3D human pose estimation in video using 2D keypoint trajectories

Unadversarial Examples: Designing Objects for Robust Vision

Tutel MoE: An Optimized Mixture-of-Experts Implementation

7th place solution of Human Protein Atlas - Single Cell Classification on Kaggle

SoK: Vehicle Orientation Representations for Deep Rotation Estimation

UCSD Oasis platform

TorchGeo is a PyTorch domain library, similar to torchvision, that provides datasets, transforms, samplers, and pre-trained models specific to geospatial data.

Some pre-commit hooks for OpenMMLab projects

SeqAttack: a framework for adversarial attacks on token classification models

Fast, flexible and fun neural networks.

A high performance implementation of HDBSCAN clustering.

A demo of how to use JAX to create a simple gravity simulation

This is a repository with the code for the ACL 2019 paper