Getting Started

Setup

Clone the repository

git clone https://github.com/avihu111/TypiClust
cd TypiClust

Create an environment using

conda create --name typiclust --file typiclust-env.txt
conda activate typiclust
pip install pyyaml easydict termcolor tqdm simplejson yacs

If that fails, it might be due to incompatible CUDA version. In that case, try installing by

conda create --name typiclust python=3.7
conda activate typiclust
conda install pytorch torchvision torchaudio cudatoolkit=<CUDA_VERSION> -c pytorch
conda install matplotlib scipy scikit-learn pandas
conda install -c conda-forge faiss-gpu
pip install pyyaml easydict termcolor tqdm simplejson yacs

Select the GPU to be used by running

CUDA_VISIBLE_DEVICES=0

Representation Learning

Both TypiClust variants and ProbCover rely on representation learning. To train CIFAR-10 on simclr please run

cd scan
python simclr.py --config_env configs/env.yml --config_exp configs/pretext/simclr_cifar10.yml
cd ..

When this finishes, the file ./results/cifar-10/pretext/features_seed1.npy should exist.

To save time, you can download the features of CIFAR-10/100 from here:

Dataset	Download link
CIFAR10	Download
CIFAR100	Download

and locate the files here:./results/cifar-10/pretext/features_seed1.npy.

TypiClust - K-Means variant

To select samples according to TypiClust (K-Means) where the initial_size=0 and the budget=100 please run

cd deep-al/tools
python train_al.py --cfg ../configs/cifar10/al/RESNET18.yaml --al typiclust_rp --exp-name auto --initial_size 0 --budget 100
cd ../../

TypiClust - SCAN variant

In this section we select budget=10 samples without an initial set. We first must run SCAN clustering algorithm, as TypiClust uses its features cluster assignments. Please repeat the following command for k in [10, 20, 30, 40, 50, 60]

cd scan
python scan.py --config_env configs/env.yml --config_exp configs/scan/scan_cifar10.yml --num_clusters k
cd .. 

You can use other representations and change the path in the file deep-al/pycls/datasets/utils/features.py. Then, you can run the active learning experiment by running

cd deep-al/tools
python train_al.py --cfg ../configs/cifar10/al/RESNET18.yaml --al typiclust_dc --exp-name auto --initial_size 0 --budget 10
cd ../../

ProbCover

Example ProbCover script

cd deep-al/tools
python train_al.py --cfg ../configs/cifar10/al/RESNET18.yaml --al probcover --exp-name auto --initial_size 0 --budget 10 --delta 0.75
cd ../../

DCoM

Example DCoM script:

cd deep-al/tools
python train_al.py --cfg ../configs/cifar10/al/RESNET18.yaml --al dcom --exp-name auto --initial_size 0 --budget 10 --initial_delta 0.75
cd ../../

You can add the a_logistic and k_logistic parameters to the run using --a_logistic and --k_logistic.