The inverse-Wishart prior for covariance matrix

Is there a prior distribution of inverse-Wishart in pymc3, I want to estimate the covariance matrix in a multivariate model, so we need to assign a inverse-Wishart prior to the covariance matrix, but not found in pymc,

Using wishart as prior for covariance matrix is a bad idea (see eg https://dahtah.wordpress.com/2012/08/22/priors-of-convenience/) - you should use LKJ (LKJCholeskyCov or LKJCorr) instead.

I need an inverse wishart priori, which is mainly used to compare with other methods,basically, it doesn’t consider the performance of it.
but I didn’t find inverse wishart in pymc3,

PyMC3 doesnt have inverse wishart implemented. If you really want you can use the Wishart distribution and inverse the random variable. But again inferencing with wishart prior is difficult so it might render the comparison meaningless.

Thank you very much for your advice,and I’ll try to find a better way.