/LADA

Official PyTorch implementation for Look-Ahead Data Acquisition via Augmentation for Deep Active Learning [NeurIPS 2021]

Primary LanguagePythonApache License 2.0Apache-2.0

LADA: Look-Ahead Data Acquisition via Augmentation for Deep Active Learning (NeurIPS 2021)

Requirements

To install requirements:

pip install -r requirements.txt

Training

To train the model(s) in the paper, run this command:

python main.py --data Cifar10 --method LADA

Evaluation

  • Data will be downloaded to folder 'data'.
  • Result will be recorded to folder 'Results'.

Results

Our model achieves the following performance on active learning settings:

Model name FashionMNIST SVHN CIFAR-10 CIFAR-100
LADA 83.68% 75.72% 53.45% 46.92%