Edit model card

swin-tiny-patch4-window7-224-dmae-va-U

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

  • Loss: 0.0900
  • Accuracy: 0.9725

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: 32
  • eval_batch_size: 32
  • seed: 42
  • gradient_accumulation_steps: 4
  • total_train_batch_size: 128
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_ratio: 0.1
  • num_epochs: 40

Training results

Training Loss Epoch Step Validation Loss Accuracy
No log 0.9 7 1.4643 0.2110
1.4218 1.94 15 1.4070 0.3303
1.3226 2.97 23 1.3454 0.3486
1.1908 4.0 31 1.1438 0.4220
1.1908 4.9 38 0.8730 0.5780
0.9441 5.94 46 0.8100 0.6422
0.7185 6.97 54 0.6099 0.7339
0.6526 8.0 62 0.5096 0.7890
0.6526 8.9 69 0.4925 0.8165
0.5185 9.94 77 0.3989 0.8349
0.4946 10.97 85 0.3276 0.8807
0.4469 12.0 93 0.3023 0.8899
0.376 12.9 100 0.3112 0.8991
0.376 13.94 108 0.2117 0.9266
0.3156 14.97 116 0.2024 0.9174
0.366 16.0 124 0.2065 0.9450
0.2806 16.9 131 0.1942 0.9174
0.2806 17.94 139 0.2393 0.9174
0.2695 18.97 147 0.1498 0.9541
0.2357 20.0 155 0.1465 0.9358
0.2345 20.9 162 0.1522 0.9633
0.2157 21.94 170 0.1403 0.9450
0.2157 22.97 178 0.0999 0.9541
0.1894 24.0 186 0.1427 0.9633
0.2195 24.9 193 0.0949 0.9633
0.1874 25.94 201 0.1152 0.9633
0.1874 26.97 209 0.1226 0.9541
0.1815 28.0 217 0.0964 0.9633
0.1619 28.9 224 0.0912 0.9633
0.201 29.94 232 0.0903 0.9633
0.1659 30.97 240 0.0745 0.9633
0.1659 32.0 248 0.0781 0.9633
0.1459 32.9 255 0.0930 0.9633
0.1459 33.94 263 0.0900 0.9725
0.1487 34.97 271 0.0796 0.9725
0.1487 36.0 279 0.0784 0.9725
0.1504 36.13 280 0.0784 0.9725

Framework versions

  • Transformers 4.35.2
  • Pytorch 2.1.0+cu118
  • Datasets 2.15.0
  • Tokenizers 0.15.0
Downloads last month
12
Safetensors
Model size
27.6M params
Tensor type
I64
·
F32
·

Finetuned from

Dataset used to train Augusto777/swin-tiny-patch4-window7-224-dmae-va-U