This is the code of "Task-Independent Knowledge Makes for Transferable Representations for Generalized Zero-Shot Learning" in AAAI2021. In this paper, we propose a novel Dual-Contrastive Embedding Network (DCEN) that simultaneously learns taskspecific and task-independent knowledge via semantic alignment and instance discrimination.

This code uses awa2 dataset.

• Install PyTorch=0.4.1.
• Install dependencies: pip install -r requirements.txt

You need to download images of animals with attributes 2 (awa2) dataset.

We adopt a two-step training strategy. Specify the image data path, save path in scripts/1.sh. Then train DCEN on one Titan XP GPU:

cd scripts
bash 1.sh

Evaluating our model (https://drive.google.com/file/d/1Mjw9kSvBpF7wcFkChMaG6uFZ9XY-7y2c/view?usp=sharing) or your own model. Specify the image data path, model path in scripts/2.sh.

cd scripts
bash 2.sh