Repository for DAMI II Project Stockholm University HT2019

David Mildenberger

Usage:

Generate challenges and prepare data sets:

python gen_ro_data.py

python -m experiments.maml2 --dataset roPUF --order 2 --n 20 --k 1 --q 1 --meta-batch-size 75 --inner-train-steps 50 --inner-val-steps 10 --inner-lr 0.01 --eval-batches 5 --epoch-len 10 --test-board D070802 --challenge-size 64 --load-indexed False --epochs 1

Experiments:

Results from paper are in results directory 64 bit model: roPUF-order=2-n=10-k=1-metabatch=60-trainsteps=40val-steps=10.zip

128 bit model: roPUF_order=2_n=90_k=1_metabatch=75_train_steps=50_val_steps=10.zip

Credits

Implementation based on https://github.com/oscarknagg/few-shot/