|
--- |
|
license: apache-2.0 |
|
base_model: facebook/deit-base-distilled-patch16-224 |
|
tags: |
|
- generated_from_trainer |
|
metrics: |
|
- precision |
|
- recall |
|
- f1 |
|
- accuracy |
|
model-index: |
|
- name: deit-cvc-drop-aug |
|
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. --> |
|
|
|
# deit-cvc-drop-aug |
|
|
|
This model is a fine-tuned version of [facebook/deit-base-distilled-patch16-224](https://huggingface.co/facebook/deit-base-distilled-patch16-224) on an unknown dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.4769 |
|
- Precision: 0.8894 |
|
- Recall: 0.8064 |
|
- F1: 0.8458 |
|
- Accuracy: 0.8489 |
|
|
|
## 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: 0.0001 |
|
- train_batch_size: 16 |
|
- eval_batch_size: 16 |
|
- seed: 17 |
|
- gradient_accumulation_steps: 6 |
|
- total_train_batch_size: 96 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: cosine |
|
- num_epochs: 15 |
|
- mixed_precision_training: Native AMP |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy | |
|
|:-------------:|:-----:|:----:|:---------------:|:---------:|:------:|:------:|:--------:| |
|
| 0.5453 | 0.27 | 100 | 0.4824 | 0.7776 | 0.7726 | 0.7751 | 0.7696 | |
|
| 0.4324 | 0.54 | 200 | 0.4796 | 0.8033 | 0.7279 | 0.7637 | 0.7686 | |
|
| 0.4042 | 0.82 | 300 | 0.3790 | 0.7697 | 0.9608 | 0.8547 | 0.8321 | |
|
| 0.3849 | 1.09 | 400 | 0.4100 | 0.8125 | 0.8198 | 0.8161 | 0.8102 | |
|
| 0.3621 | 1.36 | 500 | 0.3689 | 0.8099 | 0.8967 | 0.8511 | 0.8387 | |
|
| 0.3457 | 1.63 | 600 | 0.3313 | 0.7896 | 0.9543 | 0.8642 | 0.8459 | |
|
| 0.3443 | 1.9 | 700 | 0.3424 | 0.7836 | 0.9528 | 0.8600 | 0.8405 | |
|
| 0.3287 | 2.18 | 800 | 0.3308 | 0.8206 | 0.8947 | 0.8561 | 0.8454 | |
|
| 0.3224 | 2.45 | 900 | 0.4546 | 0.8481 | 0.6624 | 0.7438 | 0.7655 | |
|
| 0.3096 | 2.72 | 1000 | 0.3402 | 0.8300 | 0.8754 | 0.8521 | 0.8438 | |
|
| 0.3095 | 2.99 | 1100 | 0.3691 | 0.8035 | 0.9076 | 0.8524 | 0.8385 | |
|
| 0.2901 | 3.27 | 1200 | 0.3643 | 0.8008 | 0.8982 | 0.8467 | 0.8329 | |
|
| 0.2939 | 3.54 | 1300 | 0.3021 | 0.8047 | 0.9613 | 0.8760 | 0.8602 | |
|
| 0.2946 | 3.81 | 1400 | 0.3617 | 0.8363 | 0.8322 | 0.8342 | 0.8301 | |
|
| 0.2856 | 4.08 | 1500 | 0.4884 | 0.8401 | 0.7850 | 0.8116 | 0.8127 | |
|
| 0.2683 | 4.35 | 1600 | 0.3540 | 0.841 | 0.8352 | 0.8381 | 0.8341 | |
|
| 0.2724 | 4.63 | 1700 | 0.3078 | 0.8391 | 0.8957 | 0.8665 | 0.8581 | |
|
| 0.2685 | 4.9 | 1800 | 0.2913 | 0.8455 | 0.8967 | 0.8704 | 0.8627 | |
|
| 0.2449 | 5.17 | 1900 | 0.3866 | 0.8465 | 0.8515 | 0.8490 | 0.8443 | |
|
| 0.2468 | 5.44 | 2000 | 0.3072 | 0.8406 | 0.8952 | 0.8670 | 0.8589 | |
|
| 0.2557 | 5.71 | 2100 | 0.3735 | 0.8595 | 0.7900 | 0.8233 | 0.8257 | |
|
| 0.25 | 5.99 | 2200 | 0.3117 | 0.8755 | 0.8376 | 0.8561 | 0.8553 | |
|
| 0.2256 | 6.26 | 2300 | 0.3264 | 0.8407 | 0.8913 | 0.8653 | 0.8574 | |
|
| 0.234 | 6.53 | 2400 | 0.3617 | 0.8950 | 0.7572 | 0.8203 | 0.8295 | |
|
| 0.2259 | 6.8 | 2500 | 0.3284 | 0.8476 | 0.8893 | 0.8679 | 0.8609 | |
|
| 0.2261 | 7.07 | 2600 | 0.3486 | 0.9034 | 0.7805 | 0.8375 | 0.8443 | |
|
| 0.2087 | 7.35 | 2700 | 0.3971 | 0.8628 | 0.8118 | 0.8365 | 0.8369 | |
|
| 0.2035 | 7.62 | 2800 | 0.3106 | 0.8722 | 0.8942 | 0.8831 | 0.8783 | |
|
| 0.2116 | 7.89 | 2900 | 0.3734 | 0.8805 | 0.8083 | 0.8429 | 0.8451 | |
|
| 0.1956 | 8.16 | 3000 | 0.3443 | 0.8612 | 0.8654 | 0.8633 | 0.8591 | |
|
| 0.1826 | 8.44 | 3100 | 0.3795 | 0.8908 | 0.7900 | 0.8374 | 0.8423 | |
|
| 0.1918 | 8.71 | 3200 | 0.3362 | 0.8894 | 0.8267 | 0.8569 | 0.8581 | |
|
| 0.1886 | 8.98 | 3300 | 0.3259 | 0.8589 | 0.8798 | 0.8693 | 0.8640 | |
|
| 0.1716 | 9.25 | 3400 | 0.4269 | 0.8621 | 0.8347 | 0.8481 | 0.8464 | |
|
| 0.1654 | 9.52 | 3500 | 0.4066 | 0.8881 | 0.8317 | 0.8590 | 0.8597 | |
|
| 0.1625 | 9.8 | 3600 | 0.3927 | 0.8882 | 0.8128 | 0.8488 | 0.8512 | |
|
| 0.1659 | 10.07 | 3700 | 0.3797 | 0.8895 | 0.8193 | 0.8529 | 0.8548 | |
|
| 0.1519 | 10.34 | 3800 | 0.4089 | 0.8808 | 0.8217 | 0.8502 | 0.8512 | |
|
| 0.1484 | 10.61 | 3900 | 0.3865 | 0.8853 | 0.8237 | 0.8534 | 0.8546 | |
|
| 0.1427 | 10.88 | 4000 | 0.4347 | 0.8892 | 0.8009 | 0.8427 | 0.8464 | |
|
| 0.1375 | 11.16 | 4100 | 0.4688 | 0.8878 | 0.8213 | 0.8532 | 0.8548 | |
|
| 0.1276 | 11.43 | 4200 | 0.4687 | 0.8932 | 0.7974 | 0.8426 | 0.8469 | |
|
| 0.1275 | 11.7 | 4300 | 0.4493 | 0.8936 | 0.8009 | 0.8447 | 0.8487 | |
|
| 0.1349 | 11.97 | 4400 | 0.4618 | 0.8975 | 0.7825 | 0.8361 | 0.8423 | |
|
| 0.1217 | 12.24 | 4500 | 0.4636 | 0.8987 | 0.7974 | 0.8450 | 0.8497 | |
|
| 0.1211 | 12.52 | 4600 | 0.4527 | 0.8815 | 0.8307 | 0.8553 | 0.8556 | |
|
| 0.1164 | 12.79 | 4700 | 0.4669 | 0.8950 | 0.8123 | 0.8516 | 0.8546 | |
|
| 0.1119 | 13.06 | 4800 | 0.4617 | 0.8875 | 0.8148 | 0.8496 | 0.8517 | |
|
| 0.11 | 13.33 | 4900 | 0.4718 | 0.8894 | 0.8103 | 0.8480 | 0.8507 | |
|
| 0.1138 | 13.61 | 5000 | 0.4892 | 0.8939 | 0.7989 | 0.8437 | 0.8479 | |
|
| 0.1058 | 13.88 | 5100 | 0.4725 | 0.8875 | 0.8108 | 0.8474 | 0.8500 | |
|
| 0.1042 | 14.15 | 5200 | 0.4788 | 0.8908 | 0.8064 | 0.8465 | 0.8497 | |
|
| 0.107 | 14.42 | 5300 | 0.4759 | 0.8900 | 0.8073 | 0.8467 | 0.8497 | |
|
| 0.1047 | 14.69 | 5400 | 0.4767 | 0.8894 | 0.8064 | 0.8458 | 0.8489 | |
|
| 0.1085 | 14.97 | 5500 | 0.4769 | 0.8894 | 0.8064 | 0.8458 | 0.8489 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.36.2 |
|
- Pytorch 2.1.2 |
|
- Datasets 2.16.1 |
|
- Tokenizers 0.15.0 |
|
|
