| ======================================================================== |
| MODEL: cnn |
| OPTUNA objective: spearman (direction=maximize) |
| Best trial: 16 |
| Best val metrics: |
| { |
| "rmse": 1.3464045524597168, |
| "mae": 0.8111599683761597, |
| "r2": 0.18515795469284058, |
| "spearman_rho": 0.7157261392036586 |
| } |
| Saved model: /vast/projects/pranam/lab/yz927/projects/Classifier_Weight/training_classifiers/permeability_caco2/cnn_smiles/best_model.pt |
| Best params: |
| { |
| "lr": 0.000272851995406727, |
| "weight_decay": 3.0537910943840606e-05, |
| "dropout": 0.3058903047096042, |
| "batch_size": 16, |
| "channels": 128, |
| "kernel": 3, |
| "layers": 4, |
| "loss": "mse" |
| } |
| Benchmark (val): |
| { |
| "n_samples": 122, |
| "wall_time_s": 0.496349481982179, |
| "throughput_samples_per_s": 245.79455490270928, |
| "gpu_ms_per_sample": 0.030315278494944337, |
| "gpu_peak_mem_MB": 42.23388671875 |
| } |
| ======================================================================== |
