Run
conda env create -f requirements.yaml
and activate the conda environment
conda activate covmatch
Download Flickr30k and COCO datasets and put them here:
./distill_utils/data/
├── Flickr30k/
│ ├── flickr30k-images/
│ │ ├── 1234.jpg
│ │ └── ......
│ ├── results_20130124.token
│ └── readme.txt
└── COCO/
├── train2014/
├── val2014/
└── test2014/
All script files are provided in 'scripts/'. For example, to distill flickr30k with 100 synthetic pairs, run
sh scripts/distill_flickr_N100.sh
We only provide synthetic set for Flickr30k with N=100 due to space limit, in 'results/'. To evaluate synthetic sets, run
sh scripts/evaluate_all.sh