Skip to content

Latest commit

 

History

History
45 lines (37 loc) · 1.83 KB

README.md

File metadata and controls

45 lines (37 loc) · 1.83 KB

Domain Adaptation with Invariant RepresentationLearning: What Transformations to Learn?

Repository Structure:

DSAN
|└───amazon
|    └── dataset (Amazon dataset)
|    ├── result
|    ├── amazon_utils.py
|    ├── dsan.py
|    └── flip_gradient.py
|    └── logger.py
|────imageclef
|    └── dataset (ImageCLEF dataset)
|    ├── logs
|    ├── utils.py
|    ├── dsan.py
|    └── flip_gradient.py

Instructions on running the code: ##1. Run the following command



# for Amazon
cd amazon
python dsan.py --src $source_domain_name --tgt $target_domain_name 

# for ImageCLEF
cd imageclef
python dsan.py --src $source_domain_name --tgt $target_domain_name

##2. Compute environment for our experiments:
CPU: Intel 7700k
GPU: GeForce RTX2070
32 GB Memory
##3. Table of the experiment result for Amazon:

Model B→D B→E B→K D→B D→E D→K E→D E→B E→K K→B K→D K→E
DIRT-T 78.6 76.1 75.5 76.8 75.2 79.1 69.6 71.0 84.2 69.2 73.3 79.5
MDD 77.1 74.4 77.0 74.7 74.1 76.3 72.4 70.2 83.3 69.3 73.2 82.8
DSAN 82.7 80.8 82.6 79.5 81.4 85.3 76.7 75.1 88.0 73.8 77.3 85.0