Hugging Face's logo

Augusto777
/
Swin-dmae-DA3-N-Colab

TensorBoard
Safetensors
swinv2
Generated from Trainer
Eval Results
Model card Files Metrics Community

Swin-dmae-DA3-N-Colab

This model is a fine-tuned version of microsoft/swinv2-tiny-patch4-window8-256 on the imagefolder dataset. It achieves the following results on the evaluation set:

  • Loss: 1.6604
  • Accuracy: 0.7826

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: 5e-05
  • train_batch_size: 16
  • eval_batch_size: 16
  • seed: 42
  • gradient_accumulation_steps: 4
  • total_train_batch_size: 64
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_ratio: 0.1
  • num_epochs: 120

Training results

Training Loss Epoch Step Validation Loss Accuracy
1.4177 0.98 22 1.3455 0.4348
1.4043 2.0 45 1.3475 0.4565
1.3628 2.98 67 1.3397 0.4565
1.21 4.0 90 1.2593 0.4565
1.0504 4.98 112 1.1194 0.4130
0.9129 6.0 135 1.0522 0.4130
0.7811 6.98 157 1.1184 0.4348
0.6572 8.0 180 0.9951 0.5870
0.5207 8.98 202 0.9055 0.6522
0.6234 10.0 225 0.9277 0.6087
0.4721 10.98 247 0.8458 0.6739
0.3944 12.0 270 0.8837 0.6522
0.3572 12.98 292 0.8719 0.7174
0.2911 14.0 315 1.0367 0.6304
0.3224 14.98 337 0.9641 0.6087
0.2663 16.0 360 1.3670 0.5870
0.2132 16.98 382 1.3090 0.6304
0.266 18.0 405 1.1247 0.7174
0.1929 18.98 427 1.1458 0.6087
0.1831 20.0 450 1.0267 0.7391
0.2298 20.98 472 1.1863 0.6304
0.1825 22.0 495 1.0458 0.6957
0.1701 22.98 517 1.3520 0.6087
0.1964 24.0 540 1.3927 0.6522
0.1731 24.98 562 1.4361 0.6522
0.1565 26.0 585 1.0449 0.6957
0.1844 26.98 607 1.3166 0.6087
0.1187 28.0 630 1.7950 0.6304
0.129 28.98 652 1.2753 0.6957
0.1269 30.0 675 1.4244 0.6739
0.1522 30.98 697 1.4873 0.6522
0.1497 32.0 720 1.3693 0.6739
0.1215 32.98 742 1.8168 0.6739
0.1049 34.0 765 1.2749 0.7609
0.1013 34.98 787 1.5098 0.7609
0.1499 36.0 810 1.6464 0.6304
0.0823 36.98 832 1.7892 0.6957
0.092 38.0 855 1.6448 0.6739
0.1076 38.98 877 1.6955 0.6304
0.1163 40.0 900 1.5780 0.6739
0.0952 40.98 922 1.8121 0.6522
0.0833 42.0 945 1.4459 0.7174
0.1045 42.98 967 1.7307 0.6739
0.094 44.0 990 1.4970 0.7391
0.092 44.98 1012 1.5766 0.7174
0.0863 46.0 1035 1.7600 0.6522
0.101 46.98 1057 1.4763 0.6957
0.0995 48.0 1080 2.0018 0.6522
0.0893 48.98 1102 1.4872 0.7391
0.0965 50.0 1125 1.6165 0.7391
0.0595 50.98 1147 1.6608 0.7391
0.0606 52.0 1170 1.6604 0.7826
0.0794 52.98 1192 1.9967 0.7174
0.0919 54.0 1215 1.7728 0.6957
0.0666 54.98 1237 1.7364 0.7391
0.0842 56.0 1260 1.6661 0.7609
0.0781 56.98 1282 1.9340 0.7174
0.0565 58.0 1305 1.7399 0.7391
0.0939 58.98 1327 1.6644 0.7609
0.0666 60.0 1350 1.6804 0.6957
0.0577 60.98 1372 1.8968 0.6957
0.0534 62.0 1395 1.8967 0.7391
0.0592 62.98 1417 2.0113 0.7174
0.0732 64.0 1440 1.9213 0.6522
0.0768 64.98 1462 1.8912 0.7391
0.0415 66.0 1485 1.7955 0.7391
0.036 66.98 1507 1.6584 0.7391
0.0617 68.0 1530 1.9461 0.7391
0.0622 68.98 1552 1.7302 0.7826
0.0362 70.0 1575 1.7996 0.7609
0.0526 70.98 1597 1.6479 0.7391
0.0493 72.0 1620 1.7251 0.6739
0.0703 72.98 1642 2.0378 0.7391
0.0692 74.0 1665 2.0999 0.6522
0.0396 74.98 1687 2.0074 0.6739
0.0505 76.0 1710 1.7463 0.7174
0.0512 76.98 1732 1.6401 0.7609
0.0653 78.0 1755 1.7836 0.6957
0.0764 78.98 1777 1.7904 0.7609
0.0598 80.0 1800 1.8720 0.7609
0.0523 80.98 1822 1.7356 0.7174
0.0499 82.0 1845 1.9223 0.7609
0.0841 82.98 1867 1.9060 0.7609
0.0597 84.0 1890 1.9092 0.7174
0.0696 84.98 1912 1.9534 0.6957
0.0566 86.0 1935 1.9356 0.7609
0.0435 86.98 1957 2.0566 0.7391
0.024 88.0 1980 1.8717 0.7174
0.0137 88.98 2002 2.0280 0.7174
0.0663 90.0 2025 1.8829 0.7174
0.035 90.98 2047 1.9688 0.7174
0.0504 92.0 2070 2.0456 0.7391
0.025 92.98 2092 2.1568 0.6304
0.0405 94.0 2115 2.0628 0.6304
0.0247 94.98 2137 2.0686 0.6957
0.0429 96.0 2160 2.1125 0.7174
0.0408 96.98 2182 2.1003 0.7391
0.0385 98.0 2205 2.2997 0.6739
0.0364 98.98 2227 2.1442 0.7174
0.0415 100.0 2250 2.1043 0.7174
0.0175 100.98 2272 2.1847 0.6739
0.0281 102.0 2295 2.3262 0.6522
0.0268 102.98 2317 2.2843 0.6957
0.022 104.0 2340 2.3522 0.6957
0.0279 104.98 2362 2.4827 0.6739
0.0188 106.0 2385 2.5688 0.6522
0.0303 106.98 2407 2.4357 0.6522
0.0439 108.0 2430 2.4359 0.6522
0.0422 108.98 2452 2.4725 0.6739
0.032 110.0 2475 2.2899 0.6522
0.0414 110.98 2497 2.2685 0.6957
0.03 112.0 2520 2.3063 0.6739
0.0293 112.98 2542 2.3524 0.6739
0.0514 114.0 2565 2.3612 0.6739
0.0234 114.98 2587 2.3711 0.6522
0.0476 116.0 2610 2.3548 0.6739
0.0307 116.98 2632 2.3432 0.6739
0.028 117.33 2640 2.3432 0.6739

Framework versions

  • Transformers 4.36.2
  • Pytorch 2.1.2+cu118
  • Datasets 2.16.1
  • Tokenizers 0.15.0
Downloads last month
5
Safetensors
Model size
27.6M params
Tensor type
F32
·
Inference Providers NEW
This model is not currently available via any of the supported Inference Providers.
The model cannot be deployed to the HF Inference API: The model has no library tag.

Model tree for Augusto777/Swin-dmae-DA3-N-Colab

Base model

microsoft/swinv2-tiny-patch4-window8-256
Finetuned
(121)
this model

Evaluation results

View on Papers With Code