Closed BruceChaun closed 2 years ago
Summary: Merge two pretraining datasets and support round robin strategy. Each loss is computed given a certain dataset. This slightly changes the MMFLoss where it skips the computation if the current dataset name is not in the specified datasets.
MMFLoss
Reviewed By: lichengunc
Differential Revision: D33834556
This pull request was exported from Phabricator. Differential Revision: D33834556
Summary: Merge two pretraining datasets and support round robin strategy. Each loss is computed given a certain dataset. This slightly changes the
MMFLoss
where it skips the computation if the current dataset name is not in the specified datasets.Reviewed By: lichengunc
Differential Revision: D33834556