Nilearn Canonical ICA set customized fastICA iterations?

Hi Nilearn experts, I got the following warning saying need to increase iteration number when using canICA following this tutorial, I wonder is there a way to set customized fastICA #iteraiton or tolerence? From the reference, it seems there’s no parameter for setting that.

/sw/anaconda2/envs/mris/lib/python3.8/site-packages/sklearn/decomposition/ ConvergenceWarning: FastICA did not converge. Consider increasing tolerance or the maximum number of iterations.
warnings.warn('FastICA did not converge. Consider increasing ’

Some extra info: I’m running on HCP 7T minimally processed task fMRI, grayordinate surface representation, so the original data dimension is 2 (#surface coordinate * # time len). To use canICA I inserted two axis after spatial domain (i.e. made it of shape # surface coordinate * 1 * 1 * # time len)