Sampler (NUTS) proposes the same value

Then what likely happen is that the starting value is at some pathological region, which the sampler could not propose a new statue that get out from the region.

Try with other starting value (in this case using init='advi+adapt_diag' might helps).