smids_10x_deit_small_sgd_0001_fold5
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: 0.4025
- Accuracy: 0.835
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 |
|---|---|---|---|---|
| 0.999 | 1.0 | 750 | 1.0177 | 0.4867 |
| 0.9125 | 2.0 | 1500 | 0.9538 | 0.56 |
| 0.8354 | 3.0 | 2250 | 0.8848 | 0.64 |
| 0.7909 | 4.0 | 3000 | 0.8172 | 0.685 |
| 0.7315 | 5.0 | 3750 | 0.7535 | 0.7183 |
| 0.6641 | 6.0 | 4500 | 0.7023 | 0.7433 |
| 0.61 | 7.0 | 5250 | 0.6582 | 0.755 |
| 0.5883 | 8.0 | 6000 | 0.6232 | 0.7783 |
| 0.6057 | 9.0 | 6750 | 0.5936 | 0.79 |
| 0.5434 | 10.0 | 7500 | 0.5693 | 0.795 |
| 0.5298 | 11.0 | 8250 | 0.5500 | 0.7917 |
| 0.4881 | 12.0 | 9000 | 0.5324 | 0.8 |
| 0.5014 | 13.0 | 9750 | 0.5180 | 0.8 |
| 0.4862 | 14.0 | 10500 | 0.5060 | 0.8083 |
| 0.4712 | 15.0 | 11250 | 0.4949 | 0.81 |
| 0.4371 | 16.0 | 12000 | 0.4864 | 0.8117 |
| 0.4626 | 17.0 | 12750 | 0.4789 | 0.815 |
| 0.4294 | 18.0 | 13500 | 0.4706 | 0.815 |
| 0.4498 | 19.0 | 14250 | 0.4650 | 0.815 |
| 0.425 | 20.0 | 15000 | 0.4594 | 0.815 |
| 0.4212 | 21.0 | 15750 | 0.4532 | 0.8167 |
| 0.4517 | 22.0 | 16500 | 0.4489 | 0.82 |
| 0.4104 | 23.0 | 17250 | 0.4443 | 0.8167 |
| 0.4051 | 24.0 | 18000 | 0.4407 | 0.82 |
| 0.4019 | 25.0 | 18750 | 0.4371 | 0.8217 |
| 0.3884 | 26.0 | 19500 | 0.4338 | 0.825 |
| 0.3154 | 27.0 | 20250 | 0.4302 | 0.825 |
| 0.3994 | 28.0 | 21000 | 0.4273 | 0.8283 |
| 0.4061 | 29.0 | 21750 | 0.4246 | 0.83 |
| 0.4059 | 30.0 | 22500 | 0.4225 | 0.8283 |
| 0.3637 | 31.0 | 23250 | 0.4202 | 0.8267 |
| 0.3501 | 32.0 | 24000 | 0.4181 | 0.8283 |
| 0.4209 | 33.0 | 24750 | 0.4163 | 0.8317 |
| 0.3255 | 34.0 | 25500 | 0.4145 | 0.8317 |
| 0.3933 | 35.0 | 26250 | 0.4127 | 0.8317 |
| 0.3766 | 36.0 | 27000 | 0.4115 | 0.8317 |
| 0.3145 | 37.0 | 27750 | 0.4102 | 0.8317 |
| 0.3874 | 38.0 | 28500 | 0.4090 | 0.83 |
| 0.3898 | 39.0 | 29250 | 0.4079 | 0.83 |
| 0.365 | 40.0 | 30000 | 0.4069 | 0.8317 |
| 0.3728 | 41.0 | 30750 | 0.4059 | 0.8317 |
| 0.3865 | 42.0 | 31500 | 0.4051 | 0.8317 |
| 0.3813 | 43.0 | 32250 | 0.4045 | 0.8317 |
| 0.3607 | 44.0 | 33000 | 0.4040 | 0.8317 |
| 0.3955 | 45.0 | 33750 | 0.4034 | 0.8333 |
| 0.3317 | 46.0 | 34500 | 0.4031 | 0.835 |
| 0.4022 | 47.0 | 35250 | 0.4028 | 0.835 |
| 0.3888 | 48.0 | 36000 | 0.4026 | 0.835 |
| 0.3745 | 49.0 | 36750 | 0.4025 | 0.835 |
| 0.3 | 50.0 | 37500 | 0.4025 | 0.835 |
Framework versions
- Transformers 4.32.1
- Pytorch 2.1.0+cu121
- Datasets 2.12.0
- Tokenizers 0.13.2