dwavesystems / dwavesystems/dwave-pytorch-plugin
GRBM.sample is inefficient
Open
enhancement
- Dominant language
- Python
- Stars
- 39
- Forks
- 25
- Avg merge
- 28m
- Merged PRs (30d)
- 1
Description
Requiring `sampler` to be a `dimod.Sampler` makes repeated sampling over changing parameters very inefficient because you have to recreate a linear and quadratic bias dictionaries or BQMs when the GRBM's parameters have been updated.
Candidate solution: wrap `dimod.Sampler`s so it plays well with natively-pytorch samplers
Contributor guide
Assessment
This issue has not been assessed yet.