Is the calculation of covariance wrong?such as sigma = np.cov(act, rowvar=False)?

[liushanyuan18]

Hi
the output of function "get_activations_from_files" is "A numpy array of dimension (num images, 2048) ..." .
That is to say,the "act" is "A numpy array of dimension (num images, 2048) ".
In the formula "sigma = np.cov(act, rowvar=False)"，the rowvar is False.

But I think the rowvar should be True because the first dimension of "act" is "num images",so the dimension of "sigma" should be (num images,num images) instead of (2048,2048).

[Martin Heusel]

Hi,
to calculate the FID we need the covariance matrix of the activations, therefore, rowvar=False is correct.

[liushanyuan18]

Sorry, I still don’t understand , could you please tell me more?

In the numpy.cov source,I found that "rowvar : bool, optional
If rowvar is True (default), then each row represents a variable, with observations in the columns. Otherwise, the relationship is transposed: each column represents a variable, while the rows contain observations.".
In the numpy array of dimension (num images, 2048) , each row represents a variable.

[Martin Heusel]

The variables are the activations not the examples (images)

[liushanyuan18]

Thank you very much!