You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Now that we have checkpoint functionality #130, a few updates to it would be nice.
Make it easier to recover interupted samples
Currently if you interrupt sampling, you can resume it easily by rerunning the model code. It would be nice to be able to check the results thought. This is already possible, but has to be handcoded additionally: donaldRwilliams/chkptstanr#2 (comment)
I'd like this to be easier. In the case where you specify a "stop_after" argument, I can just create the brmsfit object and return it via abort https://stackoverflow.com/a/57183881/3556519
But I need to figure out if we can do that if the sampling is interrupted manually. One possibility is to add a function "get_checkpoint_results", which only takes the path, rather than needing all original arguments (I would have to add this to my fork).
The text was updated successfully, but these errors were encountered:
Now that we have checkpoint functionality #130, a few updates to it would be nice.
Currently if you interrupt sampling, you can resume it easily by rerunning the model code. It would be nice to be able to check the results thought. This is already possible, but has to be handcoded additionally: donaldRwilliams/chkptstanr#2 (comment)
I'd like this to be easier. In the case where you specify a "stop_after" argument, I can just create the brmsfit object and return it via abort https://stackoverflow.com/a/57183881/3556519
But I need to figure out if we can do that if the sampling is interrupted manually. One possibility is to add a function "get_checkpoint_results", which only takes the path, rather than needing all original arguments (I would have to add this to my fork).
The text was updated successfully, but these errors were encountered: