unable to reproduce results about molpal HOT 5 CLOSED

Can you share what the output config file contains and how you’re performing the analysis?

from molpal.

This is the generated config.ini I mentioned above.

These are what inside the chkpts folder:

None of the generated files and folders have the structures of what mentioned in the README, so I cannot really follow 'Analyzing data' section.

So I simply retrieve the top 500 scored compounds from the below file which is the same file used for objective look up and then check against the generated top_????explored_iter?.csv files to see how many top 500 compounds are recovered by the explored sets.

from molpal.

You didn’t specify your acquisition metric, so it used the default random metric. That’s why you’re observing the same results as random acquisition

from molpal.

Ok thanks, I thought it was greedy by default when I look at the minimal_config.ini

from molpal.

Oh fair point. I’ll fix the documentation for this

from molpal.