| Experiment: polyatomic+polyatomic on bindingdb |
|
|
| FOLD 1/5 |
| RMSE: 1.5744, MAE: 1.2909, R2: 0.0316 |
|
|
| FOLD 2/5 |
| RMSE: 1.6103, MAE: 1.3237, R2: 0.0171 |
|
|
| FOLD 3/5 |
| RMSE: 1.6149, MAE: 1.3295, R2: 0.0158 |
|
|
| FOLD 4/5 |
| RMSE: 1.5555, MAE: 1.2655, R2: 0.0634 |
|
|
| FOLD 5/5 |
| RMSE: 1.5907, MAE: 1.2956, R2: 0.0400 |
|
|
| ====== K-FOLD SUMMARY ====== |
| RMSE: 1.5889 [1.5701, 1.6073] |
| MAE: 1.3003 [1.2775, 1.3204] |
| R2: 0.0338 [0.0195, 0.0494] |
|
|
| ★ Using fold 4 model for test inference |
| RMSE: 1.5648, MAE: 1.2730, R2: 0.0633 |
|
|
| ====== HELD-OUT TEST METRICS ====== |
| {'rmse': 1.5647968957507463, 'mae': 1.2729694843292236, 'r2': 0.06330239772796631} |
| Test RMSE: 1.5655 (95 % CI: 1.4428–1.6954) |
|
|
| Test MAE : 1.2717 (95 % CI: 1.1765–1.3733) |
|
|
| Test R² : 0.0597 (95 % CI: -0.0280–0.1509) |
|
|
|
|
