Edit model card

swin-tiny-patch4-window7-224-finetuned-tekno24-highdata-90-2nd

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

  • Loss: 0.9708
  • Accuracy: 0.6406
  • F1: 0.6352
  • Precision: 0.6327
  • Recall: 0.6406

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: 30
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Accuracy F1 Precision Recall
1.1525 0.9908 81 1.0163 0.5300 0.5131 0.5061 0.5300
1.0727 1.9939 163 1.0486 0.5392 0.5058 0.5742 0.5392
1.0178 2.9969 245 1.0716 0.5207 0.4967 0.5015 0.5207
1.104 4.0 327 1.0413 0.5346 0.5035 0.5109 0.5346
1.126 4.9908 408 1.0294 0.5622 0.5281 0.5507 0.5622
1.1405 5.9939 490 1.0851 0.5484 0.5454 0.6048 0.5484
1.1432 6.9969 572 1.0160 0.5668 0.5442 0.5949 0.5668
1.1034 8.0 654 0.9752 0.5714 0.5517 0.5587 0.5714
1.0972 8.9908 735 0.9797 0.5714 0.5567 0.5755 0.5714
1.0691 9.9939 817 0.9662 0.5899 0.5869 0.6135 0.5899
1.1044 10.9969 899 0.9390 0.6175 0.5946 0.6186 0.6175
1.0781 12.0 981 0.9324 0.6129 0.6052 0.6175 0.6129
1.1094 12.9908 1062 0.9388 0.5853 0.5533 0.5572 0.5853
1.0778 13.9939 1144 0.9288 0.5853 0.5627 0.5752 0.5853
1.0594 14.9969 1226 0.9325 0.5853 0.5778 0.6006 0.5853
1.0903 16.0 1308 0.9501 0.6175 0.6069 0.6074 0.6175
1.0076 16.9908 1389 0.9386 0.5991 0.5876 0.5987 0.5991
1.0427 17.9939 1471 0.9403 0.5899 0.5736 0.5916 0.5899
1.0086 18.9969 1553 0.9690 0.5899 0.5813 0.5824 0.5899
1.0063 20.0 1635 0.9788 0.5760 0.5836 0.5966 0.5760
0.9835 20.9908 1716 0.9336 0.6129 0.6017 0.6138 0.6129
1.0145 21.9939 1798 0.9641 0.5945 0.5870 0.5934 0.5945
0.953 22.9969 1880 0.9869 0.6129 0.5942 0.6033 0.6129
0.9429 24.0 1962 0.9708 0.6406 0.6352 0.6327 0.6406
1.0114 24.9908 2043 0.9797 0.5760 0.5666 0.5677 0.5760
0.9788 25.9939 2125 0.9722 0.6037 0.6012 0.6047 0.6037
0.9991 26.9969 2207 0.9647 0.6221 0.6122 0.6163 0.6221
0.8959 28.0 2289 0.9737 0.6083 0.6030 0.6093 0.6083
0.9705 28.9908 2370 0.9634 0.6129 0.6077 0.6147 0.6129
0.9132 29.7248 2430 0.9626 0.6129 0.6073 0.6102 0.6129

Framework versions

  • Transformers 4.44.2
  • Pytorch 2.4.0+cu121
  • Datasets 2.21.0
  • Tokenizers 0.19.1

image/png

image/png

Downloads last month
7
Safetensors
Model size
27.6M params
Tensor type
I64
·
F32
·
Inference Examples
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.

Model tree for BTX24/swin-tiny-patch4-window7-224-finetuned-tekno24-highdata-90-2nd

Finetuned
(464)
this model