vit-msn-small-wbc-classifier-cells-separated-dataset-no-agregates-10
This model is a fine-tuned version of facebook/vit-msn-small on the imagefolder dataset. It achieves the following results on the evaluation set:
- Loss: 0.1504
- Accuracy: 0.9463
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: 4
- total_train_batch_size: 256
- 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 |
|---|---|---|---|---|
| 0.3353 | 0.9937 | 118 | 0.2299 | 0.9214 |
| 0.2884 | 1.9958 | 237 | 0.2378 | 0.9172 |
| 0.2487 | 2.9979 | 356 | 0.1871 | 0.9360 |
| 0.2347 | 4.0 | 475 | 0.1920 | 0.9328 |
| 0.2343 | 4.9937 | 593 | 0.1674 | 0.9405 |
| 0.2285 | 5.9958 | 712 | 0.1642 | 0.9426 |
| 0.2079 | 6.9979 | 831 | 0.1836 | 0.9344 |
| 0.2155 | 8.0 | 950 | 0.1661 | 0.9442 |
| 0.1954 | 8.9937 | 1068 | 0.1504 | 0.9463 |
| 0.1763 | 9.9368 | 1180 | 0.1588 | 0.9450 |
Framework versions
- Transformers 4.44.2
- Pytorch 2.4.1+cu121
- Datasets 3.2.0
- Tokenizers 0.19.1