Multi-sequence multi-step time series task

I am a college student who is new to deep learning and I have a challenging multi-sequence multi-step time series task. Currently, I have 90 days of time series dataframe, which is 90 rows, with 20 columns of features and two columns of target values, and I want to use 90 days of data to predict two columns of target values for the next four days. I would like to use the cnn network model or lstm network model in the pytorch framework, who can help me construct a baseline. I would appreciate it. If possible, please leave a message or contact my email.

This, this and this might help.