Table 2 Averaged results of the tenfold cross-validation on our MTL training set
From: Collectively encoding protein properties enriches protein language models
Model | Pre | Acc | Rec | F1 |
|---|---|---|---|---|
MT-BERT | 0.636 | 0.643 | 0.705 | 0.662 |
MT-BCNN | 0.705 | 0.767 | 0.742 | 0.723 |
MT-BLSTM | 0.731 | 0.771 | 0.756 | 0.743 |