Specifying init_sample as a SampleSet when sampling from KerberosSampler results in error about dwave-hybrid

[randomir]

Hi @asierro, I can't reproduce this error.

I have tried running examples/kerberos.py with the latest stable dwave-hybrid==0.6.6, and the latest stable dimod==0.10.17 (i.e. on a fresh install of dwave-ocean-sdk==4.5.0), and it completes as expected.

Can you please provide more details (versions of relevant packages) and a minimal code sample that reproduces the issue? Thanks!

[asierro]

import networkx as nx
import dwave_networkx as dnx
import dimod
import hybrid

G = nx.random_geometric_graph(50, radius=0.001*problem_node_count)
qubo = dnx.algorithms.independent_set.maximum_weighted_independent_set_qubo(G)
bqm = dimod.AdjVectorBQM.from_qubo(qubo)
first_result = hybrid.KerberosSampler().sample(bqm)
second_result = hybrid.KerberosSampler().sample(bqm, init_sample=first_result)

Using BQM instead of AdjVectorBQM results in a different error.

[randomir]

Yup, that's a bug! Thanks for reporting it, @asierro!

[randomir]

@asierro, the fix was released with Ocean 5.