vit-base-patch16-224-in21k-cards-june-08-cropping-filtered-preprocess-change-test-2
This model is a fine-tuned version of ansilmbabl/vit-base-patch16-224-in21k-cards-june-07-cropping-filtered-preprocess-change-test on the None dataset. It achieves the following results on the evaluation set:
- Loss: 2.5958
- Accuracy: 0.5147
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: 64
- eval_batch_size: 64
- seed: 42
- gradient_accumulation_steps: 8
- total_train_batch_size: 512
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 10
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Accuracy | Validation Loss |
|---|---|---|---|---|
| 1.0182 | 0.9998 | 1298 | 0.4287 | 1.5280 |
| 0.9583 | 1.9996 | 2596 | 0.4475 | 1.4878 |
| 0.8452 | 2.9998 | 3894 | 1.4847 | 0.4716 |
| 0.6887 | 3.9996 | 5192 | 1.5848 | 0.4736 |
| 0.5269 | 4.9994 | 6490 | 1.6689 | 0.493 |
| 0.4018 | 6.0 | 7789 | 1.8483 | 0.4986 |
| 0.2909 | 6.9998 | 9087 | 2.0319 | 0.5079 |
| 0.1823 | 7.9996 | 10385 | 2.2540 | 0.5127 |
| 0.1056 | 8.9994 | 11683 | 2.4652 | 0.511 |
| 0.0767 | 9.9985 | 12980 | 2.5958 | 0.5147 |
Framework versions
- Transformers 4.41.2
- Pytorch 2.0.1+cu117
- Datasets 2.19.2
- Tokenizers 0.19.1