<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->
vit-base-binary-isic-patch-32
This model is a fine-tuned version of google/vit-base-patch32-224-in21k on the ahishamm/isic_binary_augmented dataset. It achieves the following results on the evaluation set:
- Loss: 0.2618
- Accuracy: 0.8897
- Recall: 0.8897
- F1: 0.8897
- Precision: 0.8897
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: 0.0002
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 4
Training results
Training Loss | Epoch | Step | Validation Loss | Accuracy | Recall | F1 | Precision |
---|---|---|---|---|---|---|---|
0.4617 | 0.09 | 100 | 0.4375 | 0.8083 | 0.8083 | 0.8083 | 0.8083 |
0.28 | 0.19 | 200 | 0.3352 | 0.8575 | 0.8575 | 0.8575 | 0.8575 |
0.2684 | 0.28 | 300 | 0.3081 | 0.8521 | 0.8521 | 0.8521 | 0.8521 |
0.2755 | 0.37 | 400 | 0.3489 | 0.8433 | 0.8433 | 0.8433 | 0.8433 |
0.2683 | 0.46 | 500 | 0.3533 | 0.8120 | 0.8120 | 0.8120 | 0.8120 |
0.2013 | 0.56 | 600 | 0.3404 | 0.8683 | 0.8683 | 0.8683 | 0.8683 |
0.2892 | 0.65 | 700 | 0.2950 | 0.8688 | 0.8688 | 0.8688 | 0.8688 |
0.2326 | 0.74 | 800 | 0.2858 | 0.8875 | 0.8875 | 0.8875 | 0.8875 |
0.2642 | 0.84 | 900 | 0.3176 | 0.8666 | 0.8666 | 0.8666 | 0.8666 |
0.1676 | 0.93 | 1000 | 0.3049 | 0.8726 | 0.8726 | 0.8726 | 0.8726 |
0.1002 | 1.02 | 1100 | 0.3512 | 0.8797 | 0.8797 | 0.8797 | 0.8797 |
0.1888 | 1.12 | 1200 | 0.3286 | 0.8566 | 0.8566 | 0.8566 | 0.8566 |
0.134 | 1.21 | 1300 | 0.3178 | 0.8661 | 0.8661 | 0.8661 | 0.8661 |
0.1085 | 1.3 | 1400 | 0.2618 | 0.8897 | 0.8897 | 0.8897 | 0.8897 |
0.189 | 1.39 | 1500 | 0.3490 | 0.8737 | 0.8737 | 0.8737 | 0.8737 |
0.1781 | 1.49 | 1600 | 0.3033 | 0.8728 | 0.8728 | 0.8728 | 0.8728 |
0.0788 | 1.58 | 1700 | 0.3623 | 0.8774 | 0.8774 | 0.8774 | 0.8774 |
0.1135 | 1.67 | 1800 | 0.3341 | 0.8688 | 0.8688 | 0.8688 | 0.8688 |
0.0607 | 1.77 | 1900 | 0.2954 | 0.8874 | 0.8874 | 0.8874 | 0.8874 |
0.058 | 1.86 | 2000 | 0.3921 | 0.8780 | 0.8780 | 0.8780 | 0.8780 |
0.0873 | 1.95 | 2100 | 0.3094 | 0.8883 | 0.8883 | 0.8883 | 0.8883 |
0.0584 | 2.04 | 2200 | 0.3688 | 0.8892 | 0.8892 | 0.8892 | 0.8892 |
0.0978 | 2.14 | 2300 | 0.3874 | 0.8916 | 0.8916 | 0.8916 | 0.8916 |
0.0958 | 2.23 | 2400 | 0.3679 | 0.8927 | 0.8927 | 0.8927 | 0.8927 |
0.0559 | 2.32 | 2500 | 0.4649 | 0.8874 | 0.8874 | 0.8874 | 0.8874 |
0.0125 | 2.42 | 2600 | 0.4350 | 0.8943 | 0.8943 | 0.8943 | 0.8943 |
0.0636 | 2.51 | 2700 | 0.4195 | 0.8906 | 0.8906 | 0.8906 | 0.8906 |
0.0458 | 2.6 | 2800 | 0.4127 | 0.8980 | 0.8980 | 0.8980 | 0.8980 |
0.055 | 2.7 | 2900 | 0.4086 | 0.8993 | 0.8993 | 0.8993 | 0.8993 |
0.03 | 2.79 | 3000 | 0.3710 | 0.9013 | 0.9013 | 0.9013 | 0.9013 |
0.027 | 2.88 | 3100 | 0.4148 | 0.8940 | 0.8940 | 0.8940 | 0.8940 |
0.0384 | 2.97 | 3200 | 0.4036 | 0.8938 | 0.8938 | 0.8938 | 0.8938 |
0.0095 | 3.07 | 3300 | 0.4830 | 0.8892 | 0.8892 | 0.8892 | 0.8892 |
0.0489 | 3.16 | 3400 | 0.4590 | 0.8945 | 0.8945 | 0.8945 | 0.8945 |
0.0198 | 3.25 | 3500 | 0.4732 | 0.8972 | 0.8972 | 0.8972 | 0.8972 |
0.0604 | 3.35 | 3600 | 0.5283 | 0.8916 | 0.8916 | 0.8916 | 0.8916 |
0.0028 | 3.44 | 3700 | 0.5160 | 0.8931 | 0.8931 | 0.8931 | 0.8931 |
0.0018 | 3.53 | 3800 | 0.5038 | 0.8966 | 0.8966 | 0.8966 | 0.8966 |
0.0192 | 3.62 | 3900 | 0.5112 | 0.8975 | 0.8975 | 0.8975 | 0.8975 |
0.0455 | 3.72 | 4000 | 0.4954 | 0.8985 | 0.8985 | 0.8985 | 0.8985 |
0.0014 | 3.81 | 4100 | 0.4984 | 0.8994 | 0.8994 | 0.8994 | 0.8994 |
0.0012 | 3.9 | 4200 | 0.4987 | 0.9007 | 0.9007 | 0.9007 | 0.9007 |
0.0009 | 4.0 | 4300 | 0.5004 | 0.8998 | 0.8998 | 0.8998 | 0.8998 |
Framework versions
- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3