fix evaluation for ma models, following change in 'deeplabcut/core/metrics/api.py' by Zelin2001 · Pull Request #2872

fix evaluation for ma models, following change in 'deeplabcut/core/metrics/api.py' by Zelin2001 · Pull Request #2872 · DeepLabCut/DeepLabCut

This is a supplemental fix for Commit 40465e1 "fix - evaluation during training for ma models".

Following the principles of the previous fix, this minor change passed "compute_detection_rmse" argument to "compute_metrics()". To me, this fixed the assertion error in my multi-animal project which was comparable to what caused the 40465e1 fix.