Is it preferable for the "con" metric to have a smaller value during the optimization process?

[milkboylyf]

This is the intermediate log of my optimization problem, where I found that the loss function is：loss=1con+1(1-plddt)+rmsd，
However, in the readme document, it is stated that a higher value of the "con" metric is preferred.

365 models [0, 1, 2, 3, 4] recycles 0 hard 1 soft 0 temp 1 seqid 0.52 loss 6.15 con 5.00 plddt 0.76 ptm 0.44 i_ptm 0.29 rmsd 0.91
366 models [0, 1, 2, 3, 4] recycles 0 hard 1 soft 0 temp 1 seqid 0.52 loss 6.04 con 4.84 plddt 0.73 ptm 0.42 i_ptm 0.26 rmsd 0.93
367 models [0, 1, 2, 3, 4] recycles 0 hard 1 soft 0 temp 1 seqid 0.53 loss 6.05 con 4.71 plddt 0.70 ptm 0.38 i_ptm 0.21 rmsd 1.04
368 models [0, 1, 2, 3, 4] recycles 0 hard 1 soft 0 temp 1 seqid 0.53 loss 6.06 con 4.79 plddt 0.72 ptm 0.39 i_ptm 0.24 rmsd 0.99
369 models [0, 1, 2, 3, 4] recycles 0 hard 1 soft 0 temp 1 seqid 0.52 loss 6.07 con 4.84 plddt 0.73 ptm 0.43 i_ptm 0.28 rmsd 0.96
370 models [0, 1, 2, 3, 4] recycles 0 hard 1 soft 0 temp 1 seqid 0.52 loss 6.07 con 4.76 plddt 0.72 ptm 0.41 i_ptm 0.25 rmsd 1.04
371 models [0, 1, 2, 3, 4] recycles 0 hard 1 soft 0 temp 1 seqid 0.52 loss 5.93 con 4.65 plddt 0.71 ptm 0.39 i_ptm 0.23 rmsd 0.99
372 models [0, 1, 2, 3, 4] recycles 0 hard 1 soft 0 temp 1 seqid 0.50 loss 5.94 con 4.70 plddt 0.72 ptm 0.41 i_ptm 0.25 rmsd 0.96