Repository for the paper Noise Boosted Neural Receptive Fields
Clone this repository
git clone https://github.com/galatolofederico/noise-boosted-receptive-fields.git && cd noise-boosted-receptive-fields
Create a virtualenv and install the requirements
virtualenv --python=python3.7 env && . ./env/bin/activate
pip install -r requirements.txt
To train a model run
python train.py --model <model> --dataset <dataset>
Some arguments that you can use are
| Argument | Type | Description |
|---|---|---|
| --model | "nbrf" or "cnn" | Model to use |
| --dataset | "mnist" or"fmnist" or "kmnist" | Dataset to use |
| --noise | "dataset" or "batch" or "white" | Noise type |
| --noise-position | "head" or "backbone" | Noise position |
| --without-noise | Without the noise superclass | |
| --without-others | Without the others superclass | |
| --lr | float | Learning rate |
| --batch-size | int | Batch Size |
To run a trained model run
python interact.py --model <trained_mode> --dataset <dataset>
The code is released as Free Software under the GNU/GPLv3 license. Copying, adapting and republishing it is not only allowed but also encouraged.
For any further question feel free to reach me at federico.galatolo@ing.unipi.it or on Telegram @galatolo