Pau Farré
Pau Farré
Great work. This is missing the topology-specific `to_networkx_graph`, see https://github.com/dwavesystems/dwave-system/blob/4fa4b27a422a63afa65e90773ed4cf4d918c2f9f/dwave/system/samplers/dwave_sampler.py#L486 (ie. getting a dnx.chimera_graph with all the topology info instead of a simple nx.graph) You could just inherit that by...
Agree with Alex about the caution, but I think that the methods you're already overwriting will take care of that
>2. scaled to the problem bias range This should be the default operation in my opinion I would add 6. If the embedding is for a clique, use relative chain...
@JoelPasvolsky I meant relative to the logical problem scale. The chain strength is set to 3/2*sqrt(N)*max(abs(h), abs(J)). Then one would rescale both the chains and the problem weights in order...
Feature Request: Automatic batching of QPU calls **Motivation:** The QPU has a maximum “problem-run duration” of 1s. That means that you cannot use combinations of reads, anneal durations, etc. that...
@arcondello the error tells you how much usage time you're allowed (1000000) and how much time you are asking for, this should be enough for us to batch things up....
it also messes the sequentiality of your reads if you're doing that type of analysis
On it's own line it looks a bit more like it's part of the problem's weights. But I don't know if this could mess with the parsing in older dimod...
Getting an error with `reno`. Will try again tomorrow
Any more comments or ready to merge?