smids_5x_deit_tiny_rms_001_fold4

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: 2.2524
  • Accuracy: 0.7767

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: 0.001
  • 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
0.818 1.0 375 0.8019 0.5417
0.7912 2.0 750 0.8025 0.57
0.7276 3.0 1125 0.7672 0.6083
0.7922 4.0 1500 0.6983 0.6533
0.7335 5.0 1875 0.6685 0.6917
0.6959 6.0 2250 0.6471 0.7233
0.623 7.0 2625 0.6073 0.7233
0.6887 8.0 3000 0.6966 0.6667
0.6552 9.0 3375 0.5957 0.74
0.6126 10.0 3750 0.6205 0.7
0.5793 11.0 4125 0.5808 0.7567
0.6219 12.0 4500 0.5874 0.745
0.5436 13.0 4875 0.6140 0.7317
0.6012 14.0 5250 0.5834 0.7417
0.6043 15.0 5625 0.5539 0.75
0.5011 16.0 6000 0.5531 0.7383
0.5057 17.0 6375 0.5890 0.75
0.5517 18.0 6750 0.5510 0.7583
0.5553 19.0 7125 0.5435 0.76
0.5674 20.0 7500 0.4957 0.7933
0.4667 21.0 7875 0.5150 0.7867
0.4405 22.0 8250 0.5576 0.7867
0.4436 23.0 8625 0.4866 0.7967
0.454 24.0 9000 0.5354 0.775
0.4111 25.0 9375 0.5789 0.7717
0.4049 26.0 9750 0.5450 0.7817
0.397 27.0 10125 0.5808 0.7883
0.3436 28.0 10500 0.5933 0.7817
0.3249 29.0 10875 0.5969 0.7633
0.3897 30.0 11250 0.5739 0.7817
0.3938 31.0 11625 0.5794 0.7883
0.2714 32.0 12000 0.6582 0.775
0.2808 33.0 12375 0.6348 0.775
0.321 34.0 12750 0.7200 0.7567
0.2202 35.0 13125 0.6917 0.7817
0.1634 36.0 13500 0.7700 0.7733
0.3232 37.0 13875 0.7503 0.785
0.1845 38.0 14250 0.8724 0.7567
0.1357 39.0 14625 1.0521 0.7683
0.0994 40.0 15000 1.0716 0.77
0.0743 41.0 15375 1.1704 0.7717
0.1059 42.0 15750 1.2031 0.7783
0.0494 43.0 16125 1.3921 0.7633
0.0147 44.0 16500 1.5250 0.77
0.0663 45.0 16875 1.6538 0.7667
0.0618 46.0 17250 1.8210 0.765
0.0041 47.0 17625 1.9243 0.7617
0.0018 48.0 18000 2.1515 0.7717
0.0025 49.0 18375 2.2407 0.7683
0.0002 50.0 18750 2.2524 0.7767

Framework versions

  • Transformers 4.32.1
  • Pytorch 2.1.1+cu121
  • Datasets 2.12.0
  • Tokenizers 0.13.2