|
--- |
|
license: apache-2.0 |
|
tags: |
|
- generated_from_trainer |
|
metrics: |
|
- accuracy |
|
- f1 |
|
- precision |
|
- recall |
|
model-index: |
|
- name: vit-base-skin |
|
results: [] |
|
--- |
|
|
|
<!-- 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-skin |
|
|
|
This model is a fine-tuned version of [google/vit-base-patch16-224-in21k](https://huggingface.co/google/vit-base-patch16-224-in21k) on an unknown dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.6206 |
|
- Accuracy: 0.8705 |
|
- F1: 0.8684 |
|
- Precision: 0.8850 |
|
- Recall: 0.8705 |
|
|
|
## 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: 6 |
|
- mixed_precision_training: Native AMP |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 | Precision | Recall | |
|
|:-------------:|:-----:|:----:|:---------------:|:--------:|:------:|:---------:|:------:| |
|
| 0.8057 | 0.16 | 100 | 0.7591 | 0.7254 | 0.6902 | 0.6779 | 0.7254 | |
|
| 0.7619 | 0.32 | 200 | 0.7081 | 0.7409 | 0.6953 | 0.6920 | 0.7409 | |
|
| 0.6315 | 0.48 | 300 | 0.5954 | 0.8135 | 0.8039 | 0.8688 | 0.8135 | |
|
| 0.8311 | 0.64 | 400 | 0.5974 | 0.7927 | 0.7806 | 0.7985 | 0.7927 | |
|
| 0.5666 | 0.8 | 500 | 0.6151 | 0.7720 | 0.7727 | 0.7903 | 0.7720 | |
|
| 0.5816 | 0.96 | 600 | 0.4912 | 0.8031 | 0.7440 | 0.7008 | 0.8031 | |
|
| 0.3715 | 1.12 | 700 | 0.5772 | 0.7979 | 0.7764 | 0.8024 | 0.7979 | |
|
| 0.5411 | 1.28 | 800 | 0.5024 | 0.8342 | 0.8301 | 0.8447 | 0.8342 | |
|
| 0.474 | 1.44 | 900 | 0.4374 | 0.8342 | 0.8196 | 0.8260 | 0.8342 | |
|
| 0.4386 | 1.6 | 1000 | 0.6611 | 0.7565 | 0.7808 | 0.8456 | 0.7565 | |
|
| 0.4091 | 1.76 | 1100 | 0.5261 | 0.8031 | 0.7855 | 0.8288 | 0.8031 | |
|
| 0.4023 | 1.92 | 1200 | 0.4279 | 0.8446 | 0.8462 | 0.8687 | 0.8446 | |
|
| 0.28 | 2.08 | 1300 | 0.5927 | 0.8238 | 0.8023 | 0.8468 | 0.8238 | |
|
| 0.2408 | 2.24 | 1400 | 0.4605 | 0.8446 | 0.8399 | 0.8503 | 0.8446 | |
|
| 0.2145 | 2.4 | 1500 | 0.4865 | 0.8342 | 0.8399 | 0.8575 | 0.8342 | |
|
| 0.3194 | 2.56 | 1600 | 0.4727 | 0.8497 | 0.8435 | 0.8476 | 0.8497 | |
|
| 0.2391 | 2.72 | 1700 | 0.4676 | 0.8446 | 0.8402 | 0.8423 | 0.8446 | |
|
| 0.1828 | 2.88 | 1800 | 0.4337 | 0.8601 | 0.8625 | 0.8709 | 0.8601 | |
|
| 0.1232 | 3.04 | 1900 | 0.4549 | 0.8601 | 0.8646 | 0.8726 | 0.8601 | |
|
| 0.0929 | 3.19 | 2000 | 0.5939 | 0.8497 | 0.8521 | 0.8606 | 0.8497 | |
|
| 0.0559 | 3.35 | 2100 | 0.5807 | 0.8290 | 0.8237 | 0.8243 | 0.8290 | |
|
| 0.1833 | 3.51 | 2200 | 0.5235 | 0.8601 | 0.8610 | 0.8636 | 0.8601 | |
|
| 0.1395 | 3.67 | 2300 | 0.6750 | 0.8135 | 0.8208 | 0.8466 | 0.8135 | |
|
| 0.0485 | 3.83 | 2400 | 0.4431 | 0.8860 | 0.8856 | 0.8888 | 0.8860 | |
|
| 0.1206 | 3.99 | 2500 | 0.5491 | 0.8394 | 0.8375 | 0.8477 | 0.8394 | |
|
| 0.0485 | 4.15 | 2600 | 0.5289 | 0.8653 | 0.8677 | 0.8744 | 0.8653 | |
|
| 0.0494 | 4.31 | 2700 | 0.5665 | 0.8601 | 0.8603 | 0.8633 | 0.8601 | |
|
| 0.0062 | 4.47 | 2800 | 0.6186 | 0.8497 | 0.8479 | 0.8547 | 0.8497 | |
|
| 0.0065 | 4.63 | 2900 | 0.5823 | 0.8756 | 0.8728 | 0.8737 | 0.8756 | |
|
| 0.0045 | 4.79 | 3000 | 0.5801 | 0.8705 | 0.8699 | 0.8724 | 0.8705 | |
|
| 0.038 | 4.95 | 3100 | 0.6542 | 0.8394 | 0.8405 | 0.8472 | 0.8394 | |
|
| 0.0035 | 5.11 | 3200 | 0.6029 | 0.8653 | 0.8653 | 0.8714 | 0.8653 | |
|
| 0.0031 | 5.27 | 3300 | 0.6385 | 0.8601 | 0.8582 | 0.8653 | 0.8601 | |
|
| 0.0029 | 5.43 | 3400 | 0.6132 | 0.8705 | 0.8676 | 0.8830 | 0.8705 | |
|
| 0.0039 | 5.59 | 3500 | 0.6398 | 0.8653 | 0.8639 | 0.8815 | 0.8653 | |
|
| 0.0034 | 5.75 | 3600 | 0.6221 | 0.8653 | 0.8649 | 0.8726 | 0.8653 | |
|
| 0.003 | 5.91 | 3700 | 0.6206 | 0.8705 | 0.8684 | 0.8850 | 0.8705 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.29.2 |
|
- Pytorch 1.13.1 |
|
- Datasets 2.14.5 |
|
- Tokenizers 0.13.3 |
|
|
