/NeuralNetworkResiliency

Primary LanguageJupyter Notebook

NNResiliency

Analog Deep Neural Network Simulation

Examples

  • Train a model on the device 3 
    python main.py --net_type mlp --dataset morse --training_noise 0.05 --training_noise_mean 0 --device 3 --testing_noise 0 --testing_noise_mean 0 --batch_size 200 -a 120 --num_epochs 600
  • Test the model trained with specified parameters on the device 3. 
    python main.py --net_type mlp --dataset morse --training_noise 0.05 --training_noise_mean 0 --device 3 --testing_noise 0 0.02 0.04 0.08 0.1 0.12 0.14 0.16 0.18 0.2 --testing_noise_mean 0 --batch_size 200 -a 120 --num_epochs 600 --testOnly
    The results will be saved as json files under test/