smids_10x_deit_small_sgd_0001_fold4
This model is a fine-tuned version of facebook/deit-small-patch16-224 on the imagefolder dataset. It achieves the following results on the evaluation set:
- Loss: 0.4063
- Accuracy: 0.8417
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: 32
- eval_batch_size: 32
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 50
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 0.9715 | 1.0 | 750 | 1.0172 | 0.455 |
| 0.9076 | 2.0 | 1500 | 0.9524 | 0.5267 |
| 0.8403 | 3.0 | 2250 | 0.8812 | 0.625 |
| 0.7987 | 4.0 | 3000 | 0.8125 | 0.6817 |
| 0.7256 | 5.0 | 3750 | 0.7521 | 0.7183 |
| 0.6364 | 6.0 | 4500 | 0.7018 | 0.7483 |
| 0.5752 | 7.0 | 5250 | 0.6571 | 0.775 |
| 0.63 | 8.0 | 6000 | 0.6211 | 0.7817 |
| 0.6197 | 9.0 | 6750 | 0.5901 | 0.79 |
| 0.5118 | 10.0 | 7500 | 0.5651 | 0.7983 |
| 0.5006 | 11.0 | 8250 | 0.5449 | 0.8017 |
| 0.5617 | 12.0 | 9000 | 0.5276 | 0.8033 |
| 0.4842 | 13.0 | 9750 | 0.5134 | 0.8083 |
| 0.5031 | 14.0 | 10500 | 0.5016 | 0.81 |
| 0.4417 | 15.0 | 11250 | 0.4908 | 0.8083 |
| 0.4457 | 16.0 | 12000 | 0.4818 | 0.8083 |
| 0.3768 | 17.0 | 12750 | 0.4743 | 0.8117 |
| 0.4232 | 18.0 | 13500 | 0.4671 | 0.8167 |
| 0.4491 | 19.0 | 14250 | 0.4614 | 0.8167 |
| 0.4472 | 20.0 | 15000 | 0.4557 | 0.8233 |
| 0.3954 | 21.0 | 15750 | 0.4506 | 0.8267 |
| 0.405 | 22.0 | 16500 | 0.4463 | 0.83 |
| 0.4169 | 23.0 | 17250 | 0.4425 | 0.8317 |
| 0.4563 | 24.0 | 18000 | 0.4389 | 0.8333 |
| 0.3987 | 25.0 | 18750 | 0.4356 | 0.8333 |
| 0.39 | 26.0 | 19500 | 0.4325 | 0.8317 |
| 0.4056 | 27.0 | 20250 | 0.4297 | 0.8317 |
| 0.3872 | 28.0 | 21000 | 0.4272 | 0.8317 |
| 0.3817 | 29.0 | 21750 | 0.4249 | 0.835 |
| 0.4035 | 30.0 | 22500 | 0.4229 | 0.8367 |
| 0.3636 | 31.0 | 23250 | 0.4211 | 0.835 |
| 0.4122 | 32.0 | 24000 | 0.4193 | 0.8367 |
| 0.3917 | 33.0 | 24750 | 0.4176 | 0.8383 |
| 0.3839 | 34.0 | 25500 | 0.4161 | 0.84 |
| 0.3217 | 35.0 | 26250 | 0.4147 | 0.84 |
| 0.3641 | 36.0 | 27000 | 0.4136 | 0.84 |
| 0.3379 | 37.0 | 27750 | 0.4124 | 0.84 |
| 0.3959 | 38.0 | 28500 | 0.4115 | 0.84 |
| 0.3972 | 39.0 | 29250 | 0.4106 | 0.84 |
| 0.3899 | 40.0 | 30000 | 0.4098 | 0.84 |
| 0.3662 | 41.0 | 30750 | 0.4090 | 0.84 |
| 0.3473 | 42.0 | 31500 | 0.4084 | 0.8417 |
| 0.3905 | 43.0 | 32250 | 0.4078 | 0.8417 |
| 0.3794 | 44.0 | 33000 | 0.4074 | 0.8417 |
| 0.3783 | 45.0 | 33750 | 0.4070 | 0.8417 |
| 0.3309 | 46.0 | 34500 | 0.4067 | 0.8417 |
| 0.3086 | 47.0 | 35250 | 0.4065 | 0.8417 |
| 0.3454 | 48.0 | 36000 | 0.4063 | 0.8417 |
| 0.3559 | 49.0 | 36750 | 0.4063 | 0.8417 |
| 0.323 | 50.0 | 37500 | 0.4063 | 0.8417 |
Framework versions
- Transformers 4.32.1
- Pytorch 2.1.0+cu121
- Datasets 2.12.0
- Tokenizers 0.13.2