pm.sample() can optionally return a trace as an arviz InferenceData object instead of a MultiTrace, and soon that will be in the default behavior. InferenceData is better than MultiTrace in several ways. But it seems to be missing a key feature. When sampling in NUTS encounters divergent transition…

InferenceData is missing details about divergent transitions?

ricardoV94 March 12, 2022, 3:58pm 3

Can one of you open a request issue on our GitHub?

Topic		Replies	Views
Coordinates in PyMC & InferenceData Objects Sharing arviz	0	388	June 15, 2022
Create InferenceData from trace and sampled posterior prediction Questions	2	1586	October 12, 2020
Error occurred when I try to use InferenceData scheme Questions	1	338	October 8, 2021
Discrepancy in the total number of divergences Questions	0	535	November 5, 2019
Pm.sample result does not work with az.plot_trace or other az functions v5	9	1433	June 22, 2022