Early stopping criterion that (i) does not require a validation set, (ii) is based on a task performance metric, and (iii) allows stopping to be reached at different points for each channel.
Python source code for Channel-Wise Early Stopping without a Validation Set via NNK Polytope Interpolation
- Code has been tested for Python 3.8.
- See
requirements.txt
for required pip packages (CPU only). Please installfaiss-gpu
andtensorflow-gpu
for use with GPU.
main.py
provides an example of how to integrate CW-DeepNNK as an early stopping method in a standard ConvNet training environment in Tensorflow.run_script.bash
contains a set of example commands for using CW-DeepNNK early stopping in training and evaluating the final model.
@article{bonet2021channelwise,
title={Channel-Wise Early Stopping without a Validation Set via NNK Polytope Interpolation},
author={Bonet, David and Ortega, Antonio and Ruiz-Hidalgo, Javier and Shekkizhar, Sarath},
journal={Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC)}
year={2021},
}