| --- |
| license: apache-2.0 |
| tags: |
| - image-classification |
| - generated_from_trainer |
| metrics: |
| - accuracy |
| model-index: |
| - name: new_exper3 |
| 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. --> |
|
|
| # new_exper3 |
| |
| This model is a fine-tuned version of [google/vit-base-patch16-224-in21k](https://huggingface.co/google/vit-base-patch16-224-in21k) on the sudo-s/herbier_mesuem1 dataset. |
| It achieves the following results on the evaluation set: |
| - Loss: 0.3000 |
| - Accuracy: 0.9298 |
|
|
| ## 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: 8 |
| - seed: 42 |
| - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
| - lr_scheduler_type: linear |
| - num_epochs: 8 |
| - mixed_precision_training: Apex, opt level O1 |
|
|
| ### Training results |
|
|
| | Training Loss | Epoch | Step | Validation Loss | Accuracy | |
| |:-------------:|:-----:|:----:|:---------------:|:--------:| |
| | 4.093 | 0.16 | 100 | 4.1045 | 0.1885 | |
| | 3.5057 | 0.31 | 200 | 3.4448 | 0.3231 | |
| | 2.9116 | 0.47 | 300 | 2.9483 | 0.4537 | |
| | 2.561 | 0.63 | 400 | 2.5700 | 0.5258 | |
| | 2.1611 | 0.78 | 500 | 2.1721 | 0.6145 | |
| | 1.715 | 0.94 | 600 | 1.8255 | 0.6407 | |
| | 1.2752 | 1.1 | 700 | 1.5340 | 0.7051 | |
| | 1.2487 | 1.25 | 800 | 1.3533 | 0.7201 | |
| | 1.0333 | 1.41 | 900 | 1.1474 | 0.7826 | |
| | 0.8856 | 1.56 | 1000 | 1.0914 | 0.7645 | |
| | 0.7512 | 1.72 | 1100 | 0.8893 | 0.8119 | |
| | 0.747 | 1.88 | 1200 | 0.8370 | 0.8304 | |
| | 0.5082 | 2.03 | 1300 | 0.7131 | 0.8566 | |
| | 0.4449 | 2.19 | 1400 | 0.6573 | 0.8547 | |
| | 0.2912 | 2.35 | 1500 | 0.6184 | 0.8597 | |
| | 0.285 | 2.5 | 1600 | 0.5974 | 0.8570 | |
| | 0.2267 | 2.66 | 1700 | 0.5621 | 0.8647 | |
| | 0.2553 | 2.82 | 1800 | 0.5044 | 0.8816 | |
| | 0.2029 | 2.97 | 1900 | 0.4342 | 0.8955 | |
| | 0.1763 | 3.13 | 2000 | 0.4487 | 0.8905 | |
| | 0.1418 | 3.29 | 2100 | 0.4173 | 0.9005 | |
| | 0.0563 | 3.44 | 2200 | 0.3870 | 0.9048 | |
| | 0.0579 | 3.6 | 2300 | 0.3849 | 0.9036 | |
| | 0.166 | 3.76 | 2400 | 0.3933 | 0.9025 | |
| | 0.11 | 3.91 | 2500 | 0.3918 | 0.9056 | |
| | 0.0356 | 4.07 | 2600 | 0.3298 | 0.9202 | |
| | 0.0513 | 4.23 | 2700 | 0.3371 | 0.9210 | |
| | 0.0762 | 4.38 | 2800 | 0.3253 | 0.9225 | |
| | 0.018 | 4.54 | 2900 | 0.3467 | 0.9148 | |
| | 0.0263 | 4.69 | 3000 | 0.3544 | 0.9144 | |
| | 0.0205 | 4.85 | 3100 | 0.3340 | 0.9221 | |
| | 0.0237 | 5.01 | 3200 | 0.3353 | 0.9144 | |
| | 0.013 | 5.16 | 3300 | 0.3218 | 0.9229 | |
| | 0.0116 | 5.32 | 3400 | 0.3088 | 0.9291 | |
| | 0.0119 | 5.48 | 3500 | 0.3047 | 0.9279 | |
| | 0.0098 | 5.63 | 3600 | 0.3063 | 0.9283 | |
| | 0.0086 | 5.79 | 3700 | 0.3074 | 0.9268 | |
| | 0.0081 | 5.95 | 3800 | 0.3220 | 0.9237 | |
| | 0.0078 | 6.1 | 3900 | 0.3064 | 0.9268 | |
| | 0.0074 | 6.26 | 4000 | 0.3062 | 0.9279 | |
| | 0.0068 | 6.42 | 4100 | 0.3051 | 0.9291 | |
| | 0.006 | 6.57 | 4200 | 0.3000 | 0.9298 | |
| | 0.0075 | 6.73 | 4300 | 0.3010 | 0.9310 | |
| | 0.0057 | 6.89 | 4400 | 0.3037 | 0.9298 | |
| | 0.0058 | 7.04 | 4500 | 0.3071 | 0.9279 | |
| | 0.0075 | 7.2 | 4600 | 0.3075 | 0.9283 | |
| | 0.0066 | 7.36 | 4700 | 0.3077 | 0.9295 | |
| | 0.0056 | 7.51 | 4800 | 0.3084 | 0.9295 | |
| | 0.0053 | 7.67 | 4900 | 0.3064 | 0.9310 | |
| | 0.0057 | 7.82 | 5000 | 0.3068 | 0.9318 | |
| | 0.0055 | 7.98 | 5100 | 0.3068 | 0.9318 | |
|
|
|
|
| ### Framework versions |
|
|
| - Transformers 4.19.4 |
| - Pytorch 1.5.1 |
| - Datasets 2.3.2 |
| - Tokenizers 0.12.1 |
|
|
