When computing the effective sample size of a NumPy array as done below

import numpy as np
import tensorflow_probability as tfp

x = np.random.randn(1000, 2)
ess = tfp.mcmc.effective_sample_size(x)

I get the following cryptic warning:

WARNING:tensorflow:You are casting an input of type complex128 to an incompatible dtype float64.  This will discard the imaginary part and may not be what you intended.

How can I fix this?