How to deal with multiple features as input to a classifier with time series type data?

Question

My dataset has 10 samples and each sample contains 100 timesteps. I have 5 channels for each sample. I extracted 20 features per sample by taking the mean, median, etc. of e