gtsrb-model
This model is a fine-tuned version of google/vit-base-patch16-224-in21k on the bazyl/GTSRB dataset. It achieves the following results on the evaluation set:
- Loss: 0.0034
- Accuracy: 0.9993
Model description
The German Traffic Sign Benchmark is a multi-class, single-image classification challenge held at the International Joint Conference on Neural Networks (IJCNN) 2011. We cordially invite researchers from relevant fields to participate: The competition is designed to allow for participation without special domain knowledge. Our benchmark has the following properties:
- Single-image, multi-class classification problem
- More than 40 classes
- More than 50,000 images in total
- Large, lifelike database
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 1337
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 10.0
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 0.2593 | 1.0 | 4166 | 0.1585 | 0.9697 |
| 0.2659 | 2.0 | 8332 | 0.0472 | 0.9900 |
| 0.2825 | 3.0 | 12498 | 0.0155 | 0.9971 |
| 0.0953 | 4.0 | 16664 | 0.0113 | 0.9983 |
| 0.1277 | 5.0 | 20830 | 0.0076 | 0.9985 |
| 0.0816 | 6.0 | 24996 | 0.0047 | 0.9988 |
| 0.0382 | 7.0 | 29162 | 0.0041 | 0.9990 |
| 0.0983 | 8.0 | 33328 | 0.0059 | 0.9990 |
| 0.1746 | 9.0 | 37494 | 0.0034 | 0.9993 |
| 0.1153 | 10.0 | 41660 | 0.0038 | 0.9990 |
Framework versions
- Transformers 4.21.0.dev0
- Pytorch 1.12.0
- Datasets 2.3.2
- Tokenizers 0.12.1