GPU accelearated savgol filter in Pytorch To use: directly build or to(device) to use in GPU Current support non-batch preprocess To do: Batch support Reference: sadimanna/savitzky_golay