| Loading data... |
| Total frames: 5774 (preferred: 2920, unpreferred: 2854) |
| Train: 4619, Val: 1155 |
| Model params: 67,585 |
| Epoch 20/200: train_loss=0.2767 train_acc=0.815 val_loss=0.2894 val_acc=0.787 (best=16) |
| Epoch 40/200: train_loss=0.2738 train_acc=0.802 val_loss=0.2853 val_acc=0.783 (best=36) |
| Epoch 60/200: train_loss=0.2738 train_acc=0.798 val_loss=0.2855 val_acc=0.787 (best=58) |
| Epoch 80/200: train_loss=0.2726 train_acc=0.804 val_loss=0.2845 val_acc=0.783 (best=78) |
| Epoch 100/200: train_loss=0.2723 train_acc=0.808 val_loss=0.2838 val_acc=0.784 (best=97) |
| Epoch 120/200: train_loss=0.2726 train_acc=0.805 val_loss=0.2830 val_acc=0.786 (best=97) |
| Epoch 140/200: train_loss=0.2716 train_acc=0.809 val_loss=0.2835 val_acc=0.783 (best=97) |
| Epoch 160/200: train_loss=0.2724 train_acc=0.805 val_loss=0.2831 val_acc=0.796 (best=97) |
| Epoch 180/200: train_loss=0.2718 train_acc=0.811 val_loss=0.2839 val_acc=0.784 (best=97) |
| Epoch 200/200: train_loss=0.2718 train_acc=0.807 val_loss=0.2835 val_acc=0.780 (best=188) |
|
|
| Best epoch: 188, val_loss=0.2826 |
| Saved to /mnt/data3/maniskill3/push_cube_1pillar_dp_v2/mode_classifier/best_classifier.pt |
|
|
| === Classifier Analysis === |
| Progress 0%-5%: acc=0.506, P(pref|pref)=0.509, P(pref|unpref)=0.510 |
| Progress 5%-15%: acc=0.759, P(pref|pref)=0.769, P(pref|unpref)=0.251 |
| Progress 15%-30%: acc=1.000, P(pref|pref)=1.000, P(pref|unpref)=0.000 |
| DONE classifier |
|
|
