Code release for Learning Cross-domain Semantic-Visual Relationships for Transductive Zero-Shot Learning
Please run 'python train_TSVN.py -g <gpu_id> -i <data_dir> -o <output_dir> -m <train_dataset>' for training our model.
We upload the data and the features for label propagation:
https://drive.google.com/file/d/1QJMctqhQhgVng9an09dX0Rs9jO5Vo-Fl/view?usp=sharing
Acknowledgements: Our codes are mainly based on https://github.com/lzrobots/LearningToCompare_ZSL