Python code implementation for our proposed method, RL-OSHeDA, along with the baselines used in our experiments, is provided. We include NUSWIDE & ImageNet dataset for demonstration. Other datasets used in our experiments are publicly available and can be downloaded from the following sources:
-
ImageCLEF-DA: https://drive.google.com/drive/folders/1mxi-4MYdvC76U_eKQ_gLXY_EHvG0otHT?usp=sharing
-
CIFAR-10 & ILSVRC2012: https://drive.google.com/drive/folders/1r5Co_2rHWtVRqqre_aJlsYJ2gASFdQJM?usp=sharing
-
Office & Caltech256: https://drive.google.com/drive/folders/12iTh2DP0W3l_rtQ3ChTwx5sLsXq-Kei2?usp=share_link
-
Wikipedia: https://drive.google.com/drive/folders/1jn6ZqtqOu4ikdWwTL7YajELkVjv0YMTy?usp=sharing
-
Multilingual Reuters Collection: https://github.com/BIT-DA/SSAN/releases/download/1.0/TextToTextCategorization.zip
-
NUSWIDE & ImageNet: https://github.com/BIT-DA/SSAN/releases/download/1.0/TextToImageClassification.zip
=============================================================================================
Usage:
To run experiment for the proposed method RL-OSHeDA on NUSWIDE & ImageNet dataset, use the following command:
python main_OSHeDA.py --setting NUSIMAGE_nustag_hist_imagenet_decaf_3
To run experiment for DS3L on NUSWIDE & ImageNet dataset, use the following command:
python main_DS3L.py --setting NUSIMAGE_nustag_hist_imagenet_decaf_3
To run experiment for KPG on NUSWIDE & ImageNet dataset, use the following command:
python main_KPG.py --setting NUSIMAGE_nustag_hist_imagenet_decaf_3
To run experiment for OPDA on NUSWIDE & ImageNet dataset, use the following command:
python main_OPDA.py --setting NUSIMAGE_nustag_hist_imagenet_decaf_3
To run experiment for SCT on NUSWIDE & ImageNet dataset, use the following command:
python main_SCT.py --setting NUSIMAGE_nustag_hist_imagenet_decaf_3
To run experiment for PL on NUSWIDE & ImageNet dataset, use the following command:
python main_SPL.py --setting NUSIMAGE_nustag_hist_imagenet_decaf_3
To run experiment for SSAN on NUSWIDE & ImageNet dataset, use the following command:
python main_SSAN.py --setting NUSIMAGE_nustag_hist_imagenet_decaf_3
To run experiment for STN on NUSWIDE & ImageNet dataset, use the following command:
python main_STN.py --setting NUSIMAGE_nustag_hist_imagenet_decaf_3
To run experiment for SL on NUSWIDE & ImageNet dataset, use the following command:
python main_SUP.py --setting NUSIMAGE_nustag_hist_imagenet_decaf_3