navradio's picture
Model save
f1e309b
metadata
license: apache-2.0
base_model: microsoft/swin-tiny-patch4-window7-224
tags:
  - generated_from_trainer
datasets:
  - imagefolder
metrics:
  - accuracy
model-index:
  - name: swin-tiny-patch4-window7-224-PE
    results:
      - task:
          name: Image Classification
          type: image-classification
        dataset:
          name: imagefolder
          type: imagefolder
          config: default
          split: train
          args: default
        metrics:
          - name: Accuracy
            type: accuracy
            value: 0.797979797979798

swin-tiny-patch4-window7-224-PE

This model is a fine-tuned version of microsoft/swin-tiny-patch4-window7-224 on the imagefolder dataset. It achieves the following results on the evaluation set:

  • Loss: 0.4489
  • Accuracy: 0.7980

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: 0.0025
  • train_batch_size: 128
  • eval_batch_size: 128
  • seed: 42
  • gradient_accumulation_steps: 4
  • total_train_batch_size: 512
  • 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
0.6872 1.0 11 0.6535 0.6061
0.7287 2.0 22 0.6601 0.6397
0.7212 3.0 33 0.6740 0.5657
0.6947 4.0 44 0.6531 0.6532
0.6783 5.0 55 0.6739 0.5724
0.6816 6.0 66 0.6274 0.6599
0.6428 7.0 77 0.6671 0.6330
0.6928 8.0 88 0.6380 0.6498
0.6767 9.0 99 0.6875 0.6061
0.6918 10.0 110 0.6859 0.5690
0.6845 11.0 121 0.6810 0.5657
0.6826 12.0 132 0.6919 0.5185
0.6877 13.0 143 0.6693 0.6061
0.6709 14.0 154 0.6660 0.5690
0.6707 15.0 165 0.6764 0.5690
0.6703 16.0 176 0.6467 0.6296
0.6629 17.0 187 0.6471 0.6431
0.6557 18.0 198 0.6597 0.6229
0.659 19.0 209 0.6451 0.6027
0.65 20.0 220 0.6638 0.6094
0.6453 21.0 231 0.6544 0.6162
0.6426 22.0 242 0.6565 0.5825
0.6339 23.0 253 0.6743 0.6296
0.6236 24.0 264 0.6669 0.5960
0.6427 25.0 275 0.6379 0.6532
0.6439 26.0 286 0.6361 0.6263
0.6212 27.0 297 0.6540 0.6465
0.6186 28.0 308 0.5925 0.6700
0.6162 29.0 319 0.6224 0.6734
0.6237 30.0 330 0.6018 0.6667
0.6061 31.0 341 0.5735 0.6801
0.6138 32.0 352 0.6425 0.6566
0.595 33.0 363 0.5827 0.6768
0.5869 34.0 374 0.5956 0.7172
0.577 35.0 385 0.5458 0.7003
0.5766 36.0 396 0.5603 0.6869
0.5726 37.0 407 0.5339 0.7340
0.5702 38.0 418 0.5577 0.7138
0.5762 39.0 429 0.5262 0.7374
0.5543 40.0 440 0.5091 0.7441
0.5339 41.0 451 0.5185 0.7542
0.5428 42.0 462 0.5023 0.7542
0.5349 43.0 473 0.5439 0.7306
0.5319 44.0 484 0.4745 0.7811
0.5294 45.0 495 0.5432 0.7172
0.5314 46.0 506 0.4511 0.7912
0.5073 47.0 517 0.4379 0.8047
0.5028 48.0 528 0.4487 0.7980
0.4985 49.0 539 0.4550 0.7946
0.4826 50.0 550 0.4489 0.7980

Framework versions

  • Transformers 4.33.3
  • Pytorch 2.0.1+cu117
  • Datasets 2.14.5
  • Tokenizers 0.13.3