offline-policy-evaluation icon indicating copy to clipboard operation
offline-policy-evaluation copied to clipboard

Support running OPE on batches

Open econti opened this issue 3 years ago • 0 comments

Currently OPE is conducted on datasets that fit into memory and is done on all the data at once. We should support running OPE on batches of data and combining the results across batches.

econti avatar Jun 25 '21 02:06 econti