bilby-bot
commented
3 years ago In GitLab by @git.ligo:sylvia.biscoveanu on Mar 22, 2021, 18:09
Hello, which version of bilby are you using? The latest version of the roq_example.py should not encounter this error, since the weights are saved and loaded using weights.npz directly: https://git.ligo.org/lscsoft/bilby/-/blob/master/examples/gw_examples/injection_examples/roq_example.py#L100
For your second question, you can run using multiple cores using the npool argument for dynesty, cpnest, ptemcee, and nessai in the call to run sampler:
result = bilby.run_sampler(
likelihood=likelihood, priors=priors, sampler='dynesty', npoints=500,
injection_parameters=injection_parameters, outdir=outdir, label=label, npool=4)This will tell the sampler to run on 4 CPUs.
In GitLab by @git.ligo:support-bot on Feb 3, 2021, 04:55
Hi All, I’m using Bilby to do some work about parameter estimation on compact object with the next generation detector, such as ET or CE, when I try to run roq_example.py, One error was happened, it seems like the “weight.ison” was saved as “weight.json.npz” automtically by system and cause the OSError which shown in attachment screenshot.
Another question is, I want to plot the skymap and contour parameter about the parameter estimation, however as we know the low frequency band of CE or ET will make the numercial calculation very slowly, so besides using the ROQ method, I also use the multi-nodes to speed up the calculation, but I’m wondering the Bilby can be used in multinodes or not, I should say the multi-nodes is a trend of GW numerical calculation, or could you supply some tricks about that ? (In fact I saw one example about multi-threads with emcee, how about nestle cpnest dynesty or pymultinest ?)
Best wishes, Yutong Wang
发送自 Windows 10 版邮件应用