vit-base-patch16-224-finetuned-hateful-meme-restructured-balanced
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: 0.7145
- Accuracy: 0.556
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 |
|---|---|---|---|---|
| 0.7016 | 0.98 | 47 | 0.7243 | 0.512 |
| 0.6676 | 1.99 | 95 | 0.7139 | 0.544 |
| 0.626 | 2.99 | 143 | 0.7145 | 0.556 |
| 0.6042 | 4.0 | 191 | 0.7342 | 0.556 |
| 0.5672 | 4.98 | 238 | 0.7481 | 0.548 |
| 0.5339 | 5.99 | 286 | 0.7458 | 0.532 |
| 0.5266 | 6.99 | 334 | 0.7662 | 0.536 |
| 0.5102 | 8.0 | 382 | 0.7832 | 0.544 |
| 0.4808 | 8.98 | 429 | 0.7898 | 0.53 |
| 0.4698 | 9.84 | 470 | 0.7844 | 0.534 |
Framework versions
- Transformers 4.31.0
- Pytorch 2.0.1+cu117
- Datasets 2.13.1
- Tokenizers 0.13.3