Update Anaconda to the latest version
conda update conda
Install dependencies with Anaconda
conda env create -f environment.yml
conda activate al-env
Run active learning
python3 run.py --strategy <Strategy name> --dataset <Dataset name>
e.g to run the proposed CIRAL framework on the CIFAR-10 dataset, type the following command
python3 run.py --strategy CIRAL --dataset CIFAR10
The strategy and dataset parameter is optional and can also be specified in config.py. The different strategies and datasets are listed below. The Kaggle, AILARON and Pastore datasets needs to be separately downloaded. The Kaggle is available with the command
kaggle competitions download -c datasciencebowl
The Pastore is available at https://ibm.box.com/v/PlanktonData
| Strategy | Command |
|---|---|
| All data | ALL-DATA |
| BADGE | BADGE |
| CIRAL | CIRAL |
| CORESET | CORESET |
| DFAL | DFAL |
| K-MEANS | KMEANS |
| LEARNING LOSS | LEARNING_LOSS |
| MAX ENTROPY | MAX_ENTROPY |
| RANDOM BENCHMARK | RANDOM |
| SOFTMAX HYBRID | SOFTMAX_HYBRID |
| UNCERTAINTY | UNCERTAINTY |
| Dataset | Command |
|---|---|
| CIFAR-10 | CIFAR10 |
| Kaggle | PLANKTON10 |
| AILARON | AILARON |
| Pastore | PASTORE |
Screen session allows to detach terminal processes so that they continue when quitting the ssh-session.
to create a screen in linux type
screen
Detach screen
Ctrl+a d
List screens
screen -ls
Attach to screen
screen -r <Screen Name>
to rename the screen
Ctrl+a, :sessionname <Screen Name>
to kill a screen, must be attached to the screen
Ctrl+A K, y