Edit model card

visual_emotion_classification_vit_base_finetunned

This model is a fine-tuned version of google/vit-base-patch16-224-in21k on the imagefolder dataset. It achieves the following results on the evaluation set:

  • Loss: 1.2429
  • Accuracy: 0.5188

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: 1e-05
  • train_batch_size: 8
  • eval_batch_size: 8
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 50

Training results

Training Loss Epoch Step Validation Loss Accuracy
2.026 1.25 100 2.0071 0.275
1.8882 2.5 200 1.8921 0.3625
1.7186 3.75 300 1.7326 0.4188
1.5892 5.0 400 1.6242 0.475
1.4942 6.25 500 1.5443 0.5125
1.3825 7.5 600 1.4763 0.5062
1.3084 8.75 700 1.4554 0.4938
1.2388 10.0 800 1.4057 0.525
1.1519 11.25 900 1.3756 0.4938
1.1054 12.5 1000 1.3604 0.4875
1.0605 13.75 1100 1.3597 0.4938
1.016 15.0 1200 1.3370 0.4938
0.9601 16.25 1300 1.2981 0.4938
0.8445 17.5 1400 1.2420 0.5563
0.8514 18.75 1500 1.2485 0.5625
0.7899 20.0 1600 1.2861 0.4875
0.7459 21.25 1700 1.2860 0.4875
0.6917 22.5 1800 1.2335 0.5813
0.6864 23.75 1900 1.2726 0.5437
0.6414 25.0 2000 1.2215 0.5375
0.5583 26.25 2100 1.2756 0.5312
0.597 27.5 2200 1.2314 0.5375
0.5654 28.75 2300 1.3791 0.5125
0.5798 30.0 2400 1.1890 0.5687
0.5247 31.25 2500 1.2440 0.5687
0.5099 32.5 2600 1.2787 0.5625
0.496 33.75 2700 1.2628 0.55
0.479 35.0 2800 1.3420 0.4875
0.4685 36.25 2900 1.2817 0.5563
0.4375 37.5 3000 1.3122 0.525
0.4314 38.75 3100 1.1791 0.5563
0.4174 40.0 3200 1.2322 0.55
0.4019 41.25 3300 1.3871 0.5125
0.3738 42.5 3400 1.2854 0.5312
0.3938 43.75 3500 1.3057 0.5375
0.369 45.0 3600 1.2792 0.5437
0.3768 46.25 3700 1.2761 0.5625
0.3202 47.5 3800 1.2704 0.5375
0.3859 48.75 3900 1.2746 0.5312
0.3689 50.0 4000 1.3306 0.5563

Framework versions

  • Transformers 4.35.2
  • Pytorch 2.1.0+cu121
  • Datasets 2.17.0
  • Tokenizers 0.15.2
Downloads last month
13
Safetensors
Model size
85.8M params
Tensor type
F32
·

Finetuned from

Evaluation results