Mass matrix adaptation sampling code in notebooks fails

Open fonnesbeck opened this issue 5 years ago • 1 comments

It looks like the radon notebook was written when sample returned a tuple; it now returns an InferenceData, so it breaks. Iteration over the pm4_trace.posterior.items() also fails.

Running current master on Debian and Py 3.7.7.

Jul 06 '20 22:07 fonnesbeck

Also

radon_est = tf.gather(a, county_idx) + tf.gather(

Exception has occurred: InvalidArgumentError params must be at least 1 dimensional [Op:GatherV2]

Running current pip version on Ubuntu 20.4 and Py 3.8.2 tf-nightly-2.4

Aug 09 '20 08:08 disadone