PrateekMunjal/TorchAL

Official implementation of our paper: Towards Robust and Reproducible Active Learning using Neural Networks, accepted at CVPR 2022.

44
/ 100
Emerging

This project helps machine learning practitioners efficiently categorize large datasets by selecting the most informative data points for human annotation. It takes a large pool of unlabeled data and, through an iterative process, identifies a smaller, critical subset for manual labeling, which then trains a more accurate classification model. Data scientists and ML engineers, especially in fields with high labeling costs, would benefit from this to reduce annotation effort.

No commits in the last 6 months. Available on PyPI.

Use this if you need to train robust image classification models with deep neural networks but face high costs or time constraints for manually labeling your entire dataset.

Not ideal if your dataset is small enough to be fully labeled without significant effort, or if you are not working with image classification tasks using deep neural networks.

data-labeling image-classification dataset-optimization ML-experimentation model-training
Stale 6m
Maintenance 0 / 25
Adoption 8 / 25
Maturity 25 / 25
Community 11 / 25

How are scores calculated?

Stars

69

Forks

7

Language

Jupyter Notebook

License

MIT

Last pushed

Aug 16, 2023

Commits (30d)

0

Dependencies

12

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/PrateekMunjal/TorchAL"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.