This repository has the pytorch implementation of the paper "Generalized Zero-and Few-Shot Learning via Aligned Variational Autoencoders." (CVPR 2019) [pdf].
This repository has the implementation of zero-shot learning in a Generalized setting and has been tested on 4 datasets.
Note: I am still working on improving the results.
The dataset splits can be downloaded here, please download the Proposed Split
and place it in the same folder.
Find additional details about the dataset in the README.md
of the Proposed split
.
Download the pretrained model for various datasets [here] and place it in models/
- For Testing:
python linear_classifier.py --dataset CUB --dataset_path xlsa17/data/CUB/ --model_path models/checkpoint_cada_CUB.pth --pretrained
Change the arguments according to the dataset
- For Training:
python linear_classifier.py --dataset CUB --dataset_path xlsa17/data/CUB/
Dataset | Paper Results (s, u, h) |
Respository Results (s, u, h) |
---|---|---|
CUB | 53.5, 51.6, 52.4 | 53.52, 47.29, 50.21 |
AWA1 | 72.8, 57.3, 64.1 | 73.54, 46.69, 57.19 |
AWA2 | 75.0, 55.8, 63.9 | 82.77, 44.94, 58.25 |
SUN | 35.7, 47.2, 40.6 | 39.03, 37.43, 38.21 |