Code

Meta-Album Code Reproducibility

---

Experiments

1. Continual Learning
2. Hierarchical Classification
3. Within-Domain Few-Shot Learning
4. Cross-Domain Few-Shot Learning

Use Conda environment

Create env with python 3.8

conda create -n meta-album python=3.8

Activate env

conda activate meta-album

Get ready for experiments

cd Code

install required packages

pip install -r requirements.txt

Download data

Download data and put it in Data directory inside Code.
Data should be in Meta-Album Format

Run experiments

Experiments

Go to scripts directory

cd Scripts

Launch the bash script corresponding for the experiment you are interested in.

bash continual_learning.sh
bash hierarchical_classification.sh
bash within_domain_few_shot_learning.sh
bash cross_domain_few_shot_learning.sh

Results

The results will be stored in the following directories:

• Results/continual_learning
• Results/hierarchical_classification
• Results/within_domain_fsl
• Results/cross_domain_fsl