I am working on ECG Arrhythmia classification where each input is a ECG signal containing 6 R peak with a constant length of 4048 samples. How to optimize the kernel size? i
