Difference between MLP regressor in Sci-Kit and Pytorch

2

How different are these results?
I.e. for e.g. 10 runs what is the mean and stddev of the final result using the sklearn and PyTorch model?