https://github.com/voxel51/zcore
https://github.com/voxel51/zcore
Last synced: 12 months ago
JSON representation
- Host: GitHub
- URL: https://github.com/voxel51/zcore
- Owner: voxel51
- Created: 2024-11-15T17:54:48.000Z (over 1 year ago)
- Default Branch: main
- Last Pushed: 2025-07-01T13:32:10.000Z (about 1 year ago)
- Last Synced: 2025-07-01T14:40:48.174Z (about 1 year ago)
- Language: Shell
- Size: 3.91 MB
- Stars: 74
- Watchers: 13
- Forks: 8
- Open Issues: 1
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
- awesome-smart-sampling - Voxel51 ZCore
README
# Zero-Shot Coreset Selection
([Brent A. Griffin](https://github.com/griffbr)\*, [Jacob Marks](https://github.com/jacobmarks), [Jason J. Corso](https://github.com/jasoncorso)) @ [Voxel51](https://voxel51.com)
\* Corresponding author
**Z**ero-Shot **Core**set Selection ([ZCore](https://arxiv.org/pdf/2411.15349 "ZCore Paper")) is a method of coreset selection for unlabeled data. Deep learning methods rely on massive data, resulting in substantial costs for storage, annotation, and model training. Coreset selection aims to select a subset of the data to train models with lower cost while ideally performing on par with the full data training. Although the majority of real-world data are unlabeled, previous state-of-the-art coreset methods cannot select data that are unlabeled. As a solution, ZCore addresses the problem of coreset selection without labels _or_ training on candidate data. Instead, ZCore uses existing foundation models to generate a zero-shot embedding space for unlabeled data, then quantifies the relative importance of each example based on overall coverage and redundancy within the embedding distribution. On ImageNet, the ZCore coreset achieves a higher accuracy than previous label-based coresets at a 90% prune rate, while removing annotation requirements for 1.15 million images.
__Zero-Shot Coreset Selection Overview__
## Using ZCore
We provide example ZCore commands for coreset selection and subsequent model training for the EuroSAT10 dataset from our paper. See instructions in **Repeat Trials** to repeat experiment trials and **Dataset Setup** for full ImageNet, CIFAR, or EuroSAT setup.
Step 1. **Dataset**. [Download](https://www.dropbox.com/scl/fo/1mhwsunssr6g2v1wio0vq/AEI2cx3aZ2vWvFmSLDfUHtQ?rlkey=kbxo4uae43tnzvk6k7x5hk28u&st=8tkh3oyl&dl=0 "EuroSAT split download") and unzip ``eurosat10.zip`` in ``./data``.
Step 2. **Zero-Shot Coreset Selection**
```bash
python zeroshot_coreset_selection.py --dataset eurosat10 --data_dir ./data --results_dir ./results --embedding clip resnet18 --num_workers 10
```
[FiftyOne](https://github.com/voxel51/fiftyone) dependency to generate embeddings (``pip install fiftyone``).
Step 3. **Train Coreset Model**
```bash
python train_coreset_model.py --prune_rate 0.7 --dataset eurosat10 --data_dir ./data --score_file ./results/eurosat10/zcore-eurosat10-clip-resnet18-1000Ks-2sd-ri-1000nn-4ex-0/score.npy
```
## Repeat Trials
We provide examples scripts to repeat ZCore experiments over multiple trials in `./repeat-trial-scripts`.
Repeat **ZCore Selections** for EuroSAT10
```bash
chmod +x ./repeat-trial-scripts/eurosat10-score-x5.sh
./repeat-trial-scripts/eurosat10-score-x5.sh
```
Repeat **Coreset Model Training** for EuroSAT10
```bash
chmod +x ./repeat-trial-scripts/eurosat10-train-x5.sh
./repeat-trial-scripts/eurosat10-train-x5.sh
```
We provide example repeat trial results in `./results/example/eurosat10`. To tabulate these repeat trials run:
```bash
python process_repeat_trials.py --base_score_dir ./results/example/eurosat10/zcore-eurosat10-clip-resnet18-1000Ks-2sd-ri-1000nn-4ex
```
to generate the following table:
```
Setting p30-s51 p50-s51 p70-s51 p80-s51 p90-s51
Trial Results
0 93.80 91.93 86.10 80.98 63.63
1 93.39 91.26 85.74 78.88 65.58
2 93.63 91.21 87.91 79.84 66.70
3 93.90 92.38 86.91 79.86 65.16
4 94.06 92.26 86.47 80.20 67.75
Aggregate Results
Mean 93.76 91.81 86.63 79.95 65.76
StdDev 0.230 0.491 0.750 0.677 1.398
Overall Mean: 83.58
```
## Datasets
**ImageNet** can be downloaded [here](https://www.kaggle.com/c/imagenet-object-localization-challenge/data "ImageNet download") and subsequently reformatted using:
```bash
cd ./ILSVRC/Data/CLS-LOC/val/
wget -qO- https://raw.githubusercontent.com/soumith/imagenetloader.torch/master/valprep.sh | bash
```
**CIFAR10** and **CIFAR100** can be downloaded [here](https://www.cs.toronto.edu/~kriz/cifar.html "CIFAR download").
**EuroSAT80**, **EuroSAT40**, **EuroSAT20**, and **EuroSAT10** can be downloaded [here](https://www.dropbox.com/scl/fo/1mhwsunssr6g2v1wio0vq/AEI2cx3aZ2vWvFmSLDfUHtQ?rlkey=kbxo4uae43tnzvk6k7x5hk28u&st=8tkh3oyl&dl=0 "EuroSAT split download").
## Citation
If you find this code useful, please consider citing our [paper](https://arxiv.org/pdf/2411.15349):
```bibtex
@article{griffin24zcore,
title={Zero-Shot Coreset Selection: Efficient Pruning for Unlabeled Data},
author={Griffin, Brent A and Marks, Jacob and Corso, Jason J},
journal={arXiv preprint arXiv:2411.15349},
year={2024}
}
```
You may also want to check out our open-source toolkit, [FiftyOne](https://voxel51.com/fiftyone), which provides a powerful interface for exploring, analyzing, and visualizing datasets for computer vision and machine learning.