vit-base-patch16-224-in21k-cifar100-13
This model is a fine-tuned version of google/vit-base-patch16-224-in21k on the cifar100 dataset. It achieves the following results on the evaluation set:
- Loss: 0.3356
- Accuracy: 0.9196
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: 2e-05
- train_batch_size: 32
- eval_batch_size: 256
- seed: 13
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 10
Training results
Training Loss | Epoch | Step | Validation Loss | Accuracy |
---|---|---|---|---|
3.8097 | 1.0 | 1329 | 2.5052 | 0.8005 |
2.5636 | 2.0 | 2658 | 1.4118 | 0.8727 |
1.8249 | 3.0 | 3987 | 0.8596 | 0.8953 |
1.0627 | 4.0 | 5316 | 0.6014 | 0.9061 |
0.9053 | 5.0 | 6645 | 0.4856 | 0.9076 |
0.8036 | 6.0 | 7974 | 0.4101 | 0.9133 |
0.6851 | 7.0 | 9303 | 0.3778 | 0.9147 |
0.65 | 8.0 | 10632 | 0.3522 | 0.9179 |
0.6105 | 9.0 | 11961 | 0.3401 | 0.9189 |
0.5813 | 10.0 | 13290 | 0.3356 | 0.9196 |
Framework versions
- Transformers 4.28.0
- Pytorch 2.1.0
- Datasets 2.18.0
- Tokenizers 0.13.3
- Downloads last month
- 3