Dear Pe'er lab,
Many thanks for the great tool. I am writing to request whether it might be possible to consider allowing spectra to run the model in an interrupted manner?
As the datasets I am working on are quite large, my HPC service will not allow indefinite runs and the recommended 10,000 would take an estimated 20 days.
I see a GPU accelerated version is in development, but for now, is there a possibility to update models with more epochs?
Dear Pe'er lab,
Many thanks for the great tool. I am writing to request whether it might be possible to consider allowing spectra to run the model in an interrupted manner?
As the datasets I am working on are quite large, my HPC service will not allow indefinite runs and the recommended 10,000 would take an estimated 20 days.
I see a GPU accelerated version is in development, but for now, is there a possibility to update models with more epochs?