Hi all,
I have a deep pick model that I’m happy with and I’ve started inference but it seems very slow (likely to take over 50 hours). There isn’t an option to parallelize in the builder; has this functionality been removed?
Cheers,
Fred
@sheff_diamond_em Parallelization has been disabled for this job type due to concerns about the correctness of results. As a workaround, may we suggest to split the set of exposures into batches using the Exposure Sets Tool and create a Deep Picker Inference job for each batch. The jobs’ output particles can be re-combined by connecting them simultaneously to a downstream job.