metadata
license: apache-2.0
library_name: peft
tags:
- generated_from_trainer
base_model: google/vit-base-patch16-224-in21k
datasets:
- medmnist-v2
metrics:
- accuracy
- precision
- recall
- f1
model-index:
- name: pneumoniamnist-vit-base-finetuned
results: []
pneumoniamnist-vit-base-finetuned
This model is a fine-tuned version of google/vit-base-patch16-224-in21k on the medmnist-v2 dataset. It achieves the following results on the evaluation set:
- Loss: 0.3312
- Accuracy: 0.8878
- Precision: 0.9217
- Recall: 0.8513
- F1: 0.8712
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: 0.005
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 64
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 10
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Precision | Recall | F1 |
|---|---|---|---|---|---|---|---|
| 0.2447 | 0.9898 | 73 | 0.1180 | 0.9561 | 0.9313 | 0.9608 | 0.9446 |
| 0.2136 | 1.9932 | 147 | 0.1015 | 0.9637 | 0.9498 | 0.9562 | 0.9529 |
| 0.1431 | 2.9966 | 221 | 0.0729 | 0.9752 | 0.9732 | 0.9615 | 0.9672 |
| 0.1576 | 4.0 | 295 | 0.0873 | 0.9637 | 0.9480 | 0.9586 | 0.9532 |
| 0.2072 | 4.9898 | 368 | 0.0761 | 0.9714 | 0.9616 | 0.9638 | 0.9627 |
| 0.1908 | 5.9932 | 442 | 0.1044 | 0.9599 | 0.9348 | 0.9682 | 0.9496 |
| 0.1637 | 6.9966 | 516 | 0.0742 | 0.9676 | 0.9512 | 0.9661 | 0.9583 |
| 0.1385 | 8.0 | 590 | 0.1843 | 0.9313 | 0.8947 | 0.9537 | 0.9169 |
| 0.1335 | 8.9898 | 663 | 0.0677 | 0.9752 | 0.9626 | 0.9736 | 0.9680 |
| 0.1186 | 9.8983 | 730 | 0.0765 | 0.9752 | 0.9626 | 0.9736 | 0.9680 |
Framework versions
- PEFT 0.11.1
- Transformers 4.41.1
- Pytorch 2.3.0+cu121
- Datasets 2.19.1
- Tokenizers 0.19.1