vit-base-patch16-224-finetuned-hateful-meme-restructured
This model is a fine-tuned version of google/vit-base-patch16-224 on the imagefolder dataset. It achieves the following results on the evaluation set:
- Loss: 0.7152
- Accuracy: 0.552
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: 32
- eval_batch_size: 32
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 128
- 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
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 0.6546 | 0.99 | 66 | 0.7185 | 0.52 |
| 0.6222 | 2.0 | 133 | 0.7152 | 0.552 |
| 0.5986 | 2.99 | 199 | 0.7344 | 0.542 |
| 0.5535 | 4.0 | 266 | 0.7782 | 0.514 |
| 0.5377 | 4.99 | 332 | 0.8329 | 0.514 |
| 0.5115 | 6.0 | 399 | 0.7596 | 0.528 |
| 0.5133 | 6.99 | 465 | 0.8151 | 0.512 |
| 0.511 | 8.0 | 532 | 0.7897 | 0.538 |
| 0.4712 | 8.99 | 598 | 0.8539 | 0.514 |
| 0.4626 | 9.92 | 660 | 0.8449 | 0.522 |
Framework versions
- Transformers 4.31.0
- Pytorch 2.0.1+cu117
- Datasets 2.13.1
- Tokenizers 0.13.3