vit-base-patch16-224-in21k_Bart_or_Homer
This model is a fine-tuned version of google/vit-base-patch16-224-in21k. It achieves the following results on the evaluation set:
- Loss: 0.0636
- Accuracy: 0.9863
- F1: 0.9841
- Recall: 1.0
- Precision: 0.9688
Model description
This is a binary classification model to distinguish between Bart and Homer Simpson.
For more information on how it was created, check out the following link:https://github.com/DunnBC22/Vision_Audio_and_Multimodal_Projects/blob/main/Computer%20Vision/Image%20Classification/Binary%20Classification/Bart%20vs%20Homer/Bart_vs_Homer_Image_clf_ViT.ipynb
Intended uses & limitations
This model is intended to demonstrate my ability to solve a complex problem using technology.
Training and evaluation data
Dataset Source: https://www.kaggle.com/datasets/williamu32/dataset-bart-or-homer
Sample Images From Dataset:
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.0002
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 | Recall | Precision |
|---|---|---|---|---|---|---|---|
| 0.6996 | 1.0 | 13 | 0.1327 | 0.9726 | 0.9688 | 1.0 | 0.9394 |
| 0.6996 | 2.0 | 26 | 0.0636 | 0.9863 | 0.9841 | 1.0 | 0.9688 |
| 0.6996 | 3.0 | 39 | 0.1420 | 0.9452 | 0.9394 | 1.0 | 0.8857 |
Framework versions
- Transformers 4.25.1
- Pytorch 1.12.1
- Datasets 2.4.0
- Tokenizers 0.12.1
