--- license: apache-2.0 base_model: facebook/dinov2-large tags: - generated_from_trainer model-index: - name: bd_ortho-DinoVdeau-large-2024_11_27-batch-size64_freeze_probs results: [] --- # bd_ortho-DinoVdeau-large-2024_11_27-batch-size64_freeze_probs This model is a fine-tuned version of [facebook/dinov2-large](https://huggingface.co/facebook/dinov2-large) on the None dataset. It achieves the following results on the evaluation set: - Loss: 0.4551 - Rmse: 0.0866 - Mae: 0.0630 - Kl Divergence: 0.1147 - Explained Variance: 0.6593 - Learning Rate: 0.0000 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.001 - train_batch_size: 64 - eval_batch_size: 64 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 150 - mixed_precision_training: Native AMP ### Training results | Training Loss | Epoch | Step | Validation Loss | Rmse | Mae | Kl Divergence | Explained Variance | Rate | |:-------------:|:-----:|:-----:|:---------------:|:------:|:------:|:-------------:|:------------------:|:------:| | No log | 1.0 | 221 | 0.4634 | 0.1018 | 0.0760 | 0.0696 | 0.5492 | 0.001 | | No log | 2.0 | 442 | 0.4593 | 0.0952 | 0.0716 | 0.0038 | 0.6113 | 0.001 | | 0.5185 | 3.0 | 663 | 0.4574 | 0.0918 | 0.0670 | 0.0583 | 0.6245 | 0.001 | | 0.5185 | 4.0 | 884 | 0.4595 | 0.0955 | 0.0713 | -0.0650 | 0.6130 | 0.001 | | 0.4806 | 5.0 | 1105 | 0.4593 | 0.0954 | 0.0702 | -0.0835 | 0.6206 | 0.001 | | 0.4806 | 6.0 | 1326 | 0.4608 | 0.0977 | 0.0728 | -0.0705 | 0.6041 | 0.001 | | 0.4786 | 7.0 | 1547 | 0.4581 | 0.0927 | 0.0683 | -0.0044 | 0.6283 | 0.001 | | 0.4786 | 8.0 | 1768 | 0.4573 | 0.0916 | 0.0680 | 0.0799 | 0.6277 | 0.001 | | 0.4786 | 9.0 | 1989 | 0.4594 | 0.0947 | 0.0706 | 0.0233 | 0.6196 | 0.001 | | 0.4776 | 10.0 | 2210 | 0.4577 | 0.0918 | 0.0675 | 0.0885 | 0.6293 | 0.001 | | 0.4776 | 11.0 | 2431 | 0.4564 | 0.0898 | 0.0662 | 0.1296 | 0.6422 | 0.001 | | 0.4772 | 12.0 | 2652 | 0.4572 | 0.0913 | 0.0677 | -0.0061 | 0.6386 | 0.001 | | 0.4772 | 13.0 | 2873 | 0.4623 | 0.1002 | 0.0747 | -0.2060 | 0.6186 | 0.001 | | 0.4769 | 14.0 | 3094 | 0.4578 | 0.0925 | 0.0678 | -0.0371 | 0.6346 | 0.001 | | 0.4769 | 15.0 | 3315 | 0.4575 | 0.0917 | 0.0667 | 0.0458 | 0.6340 | 0.001 | | 0.4766 | 16.0 | 3536 | 0.4579 | 0.0926 | 0.0680 | 0.0151 | 0.6277 | 0.001 | | 0.4766 | 17.0 | 3757 | 0.4592 | 0.0949 | 0.0702 | -0.0679 | 0.6246 | 0.001 | | 0.4766 | 18.0 | 3978 | 0.4557 | 0.0887 | 0.0651 | 0.0421 | 0.6493 | 0.0001 | | 0.4758 | 19.0 | 4199 | 0.4556 | 0.0885 | 0.0647 | 0.0468 | 0.6508 | 0.0001 | | 0.4758 | 20.0 | 4420 | 0.4555 | 0.0884 | 0.0648 | 0.0405 | 0.6518 | 0.0001 | | 0.4741 | 21.0 | 4641 | 0.4555 | 0.0884 | 0.0650 | 0.0475 | 0.6533 | 0.0001 | | 0.4741 | 22.0 | 4862 | 0.4555 | 0.0883 | 0.0646 | 0.0570 | 0.6535 | 0.0001 | | 0.4738 | 23.0 | 5083 | 0.4551 | 0.0874 | 0.0641 | 0.0887 | 0.6570 | 0.0001 | | 0.4738 | 24.0 | 5304 | 0.4552 | 0.0878 | 0.0642 | 0.0555 | 0.6553 | 0.0001 | | 0.4736 | 25.0 | 5525 | 0.4552 | 0.0878 | 0.0645 | 0.0238 | 0.6582 | 0.0001 | | 0.4736 | 26.0 | 5746 | 0.4557 | 0.0885 | 0.0646 | 0.0409 | 0.6572 | 0.0001 | | 0.4736 | 27.0 | 5967 | 0.4551 | 0.0876 | 0.0639 | 0.0548 | 0.6576 | 0.0001 | | 0.4731 | 28.0 | 6188 | 0.4551 | 0.0876 | 0.0642 | 0.0273 | 0.6588 | 0.0001 | | 0.4731 | 29.0 | 6409 | 0.4548 | 0.0869 | 0.0634 | 0.0744 | 0.6618 | 0.0001 | | 0.4727 | 30.0 | 6630 | 0.4549 | 0.0873 | 0.0636 | 0.0492 | 0.6595 | 0.0001 | | 0.4727 | 31.0 | 6851 | 0.4548 | 0.0869 | 0.0632 | 0.0688 | 0.6613 | 0.0001 | | 0.4732 | 32.0 | 7072 | 0.4550 | 0.0874 | 0.0639 | 0.0271 | 0.6602 | 0.0001 | | 0.4732 | 33.0 | 7293 | 0.4554 | 0.0882 | 0.0647 | -0.0174 | 0.6580 | 0.0001 | | 0.4725 | 34.0 | 7514 | 0.4546 | 0.0866 | 0.0628 | 0.1094 | 0.6616 | 0.0001 | | 0.4725 | 35.0 | 7735 | 0.4550 | 0.0874 | 0.0639 | 0.0571 | 0.6583 | 0.0001 | | 0.4725 | 36.0 | 7956 | 0.4548 | 0.0869 | 0.0629 | 0.1453 | 0.6616 | 0.0001 | | 0.4727 | 37.0 | 8177 | 0.4553 | 0.0881 | 0.0645 | -0.0152 | 0.6587 | 0.0001 | | 0.4727 | 38.0 | 8398 | 0.4548 | 0.0870 | 0.0636 | 0.0490 | 0.6613 | 0.0001 | | 0.4727 | 39.0 | 8619 | 0.4548 | 0.0870 | 0.0631 | 0.0726 | 0.6610 | 0.0001 | | 0.4727 | 40.0 | 8840 | 0.4548 | 0.0870 | 0.0632 | 0.0637 | 0.6605 | 0.0001 | | 0.4721 | 41.0 | 9061 | 0.4547 | 0.0869 | 0.0634 | 0.0390 | 0.6628 | 1e-05 | | 0.4721 | 42.0 | 9282 | 0.4544 | 0.0862 | 0.0628 | 0.1115 | 0.6657 | 1e-05 | | 0.4721 | 43.0 | 9503 | 0.4546 | 0.0866 | 0.0632 | 0.0533 | 0.6646 | 1e-05 | | 0.4721 | 44.0 | 9724 | 0.4545 | 0.0864 | 0.0625 | 0.1350 | 0.6648 | 1e-05 | | 0.4721 | 45.0 | 9945 | 0.4550 | 0.0874 | 0.0642 | 0.0044 | 0.6625 | 1e-05 | | 0.4716 | 46.0 | 10166 | 0.4546 | 0.0867 | 0.0632 | 0.0389 | 0.6642 | 1e-05 | | 0.4716 | 47.0 | 10387 | 0.4545 | 0.0866 | 0.0630 | 0.0370 | 0.6651 | 1e-05 | | 0.4722 | 48.0 | 10608 | 0.4546 | 0.0868 | 0.0634 | 0.0194 | 0.6645 | 1e-05 | | 0.4722 | 49.0 | 10829 | 0.4544 | 0.0862 | 0.0627 | 0.0667 | 0.6667 | 0.0000 | | 0.4717 | 50.0 | 11050 | 0.4545 | 0.0865 | 0.0631 | 0.0548 | 0.6651 | 0.0000 | | 0.4717 | 51.0 | 11271 | 0.4545 | 0.0865 | 0.0629 | 0.0428 | 0.6651 | 0.0000 | | 0.4717 | 52.0 | 11492 | 0.4542 | 0.0859 | 0.0623 | 0.1236 | 0.6672 | 0.0000 | | 0.4718 | 53.0 | 11713 | 0.4542 | 0.0859 | 0.0625 | 0.0887 | 0.6672 | 0.0000 | | 0.4718 | 54.0 | 11934 | 0.4543 | 0.0862 | 0.0624 | 0.0917 | 0.6653 | 0.0000 | | 0.4716 | 55.0 | 12155 | 0.4546 | 0.0865 | 0.0631 | 0.0774 | 0.6650 | 0.0000 | | 0.4716 | 56.0 | 12376 | 0.4546 | 0.0866 | 0.0633 | 0.0473 | 0.6649 | 0.0000 | | 0.4717 | 57.0 | 12597 | 0.4549 | 0.0871 | 0.0639 | -0.0046 | 0.6658 | 0.0000 | | 0.4717 | 58.0 | 12818 | 0.4544 | 0.0864 | 0.0627 | 0.0553 | 0.6656 | 0.0000 | | 0.4716 | 59.0 | 13039 | 0.4545 | 0.0865 | 0.0631 | 0.0368 | 0.6654 | 0.0000 | | 0.4716 | 60.0 | 13260 | 0.4544 | 0.0863 | 0.0629 | 0.0471 | 0.6660 | 0.0000 | | 0.4716 | 61.0 | 13481 | 0.4542 | 0.0860 | 0.0624 | 0.0928 | 0.6670 | 0.0000 | | 0.4718 | 62.0 | 13702 | 0.4545 | 0.0866 | 0.0632 | 0.0286 | 0.6661 | 0.0000 | ### Framework versions - Transformers 4.41.0 - Pytorch 2.5.0+cu124 - Datasets 3.0.2 - Tokenizers 0.19.1