Edit model card

hushem_40x_deit_tiny_rms_00001_fold3

This model is a fine-tuned version of facebook/deit-tiny-patch16-224 on the imagefolder dataset. It achieves the following results on the evaluation set:

  • Loss: 0.4488
  • Accuracy: 0.9302

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: 1e-05
  • 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
0.0843 1.0 217 0.4280 0.8837
0.0143 2.0 434 0.2889 0.9302
0.0172 3.0 651 0.5423 0.9070
0.0189 4.0 868 1.1419 0.7907
0.0003 5.0 1085 0.4120 0.9302
0.0 6.0 1302 0.4870 0.9302
0.0 7.0 1519 0.5568 0.9070
0.0 8.0 1736 0.5757 0.8837
0.0 9.0 1953 0.6076 0.8837
0.0 10.0 2170 0.6516 0.8837
0.0 11.0 2387 0.6056 0.8837
0.0 12.0 2604 0.6691 0.8837
0.0 13.0 2821 0.6559 0.8837
0.0 14.0 3038 0.7098 0.9070
0.0 15.0 3255 0.6515 0.9070
0.0157 16.0 3472 0.6215 0.8837
0.0 17.0 3689 0.6307 0.8837
0.0 18.0 3906 0.7467 0.8837
0.0 19.0 4123 0.7677 0.8837
0.0 20.0 4340 0.7998 0.8605
0.0 21.0 4557 0.8197 0.8605
0.0 22.0 4774 0.8507 0.8605
0.0 23.0 4991 0.8634 0.8605
0.0 24.0 5208 0.8853 0.8605
0.0 25.0 5425 0.7783 0.9070
0.0 26.0 5642 0.7092 0.9302
0.0 27.0 5859 0.6309 0.9302
0.0 28.0 6076 0.6509 0.9302
0.0 29.0 6293 0.5569 0.9070
0.0 30.0 6510 0.5554 0.9302
0.0 31.0 6727 0.5595 0.9070
0.0 32.0 6944 0.5154 0.9302
0.0 33.0 7161 0.5043 0.9070
0.0 34.0 7378 0.5110 0.9535
0.0 35.0 7595 0.4416 0.9302
0.0 36.0 7812 0.4610 0.9535
0.0 37.0 8029 0.5159 0.9302
0.0 38.0 8246 0.5232 0.9302
0.0 39.0 8463 0.5109 0.9302
0.0 40.0 8680 0.4511 0.9535
0.0 41.0 8897 0.4620 0.9302
0.0 42.0 9114 0.4370 0.9302
0.0 43.0 9331 0.4660 0.9302
0.0 44.0 9548 0.4561 0.9302
0.0 45.0 9765 0.4386 0.9302
0.0 46.0 9982 0.4625 0.9302
0.0 47.0 10199 0.4505 0.9302
0.0 48.0 10416 0.4377 0.9302
0.0 49.0 10633 0.4484 0.9302
0.0 50.0 10850 0.4488 0.9302

Framework versions

  • Transformers 4.32.1
  • Pytorch 2.1.1+cu121
  • Datasets 2.12.0
  • Tokenizers 0.13.2
Downloads last month
14

Finetuned from

Evaluation results