Provide the functionality in the model_fn in run_model.py to calculate the loss when multiple labels are provided. This is a prerequisite for multi-task learning. The method for combining the losses for the different tasks could be configurable. This is best done after #117 is merged.
aribrill
commented
3 years ago
Provide the functionality in the model_fn in run_model.py to calculate the loss when multiple labels are provided. This is a prerequisite for multi-task learning. The method for combining the losses for the different tasks could be configurable. This is best done after #117 is merged.