---
library_name: transformers
license: apache-2.0
base_model: google/vit-base-patch16-224
tags:
- generated_from_trainer
datasets:
- webdataset
metrics:
- accuracy
- f1
- precision
- recall
model-index:
- name: frost-vision-v2-google_vit-base-patch16-224-v2024-11-09
  results:
  - task:
      name: Image Classification
      type: image-classification
    dataset:
      name: webdataset
      type: webdataset
      config: default
      split: train
      args: default
    metrics:
    - name: Accuracy
      type: accuracy
      value: 0.9316901408450704
    - name: F1
      type: f1
      value: 0.8267857142857142
    - name: Precision
      type: precision
      value: 0.8312387791741472
    - name: Recall
      type: recall
      value: 0.822380106571936
---

<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->

# frost-vision-v2-google_vit-base-patch16-224-v2024-11-09

This model is a fine-tuned version of [google/vit-base-patch16-224](https://huggingface.co/google/vit-base-patch16-224) on the webdataset dataset.
It achieves the following results on the evaluation set:
- Loss: 0.1769
- Accuracy: 0.9317
- F1: 0.8268
- Precision: 0.8312
- Recall: 0.8224

## 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: 0.0002
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 30
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch   | Step | Validation Loss | Accuracy | F1     | Precision | Recall |
|:-------------:|:-------:|:----:|:---------------:|:--------:|:------:|:---------:|:------:|
| 0.2067        | 1.4085  | 100  | 0.2229          | 0.9155   | 0.7736 | 0.8249    | 0.7282 |
| 0.1989        | 2.8169  | 200  | 0.2252          | 0.9102   | 0.7650 | 0.7950    | 0.7371 |
| 0.1364        | 4.2254  | 300  | 0.1834          | 0.9268   | 0.8163 | 0.8120    | 0.8206 |
| 0.1368        | 5.6338  | 400  | 0.1874          | 0.9268   | 0.7981 | 0.8801    | 0.7300 |
| 0.1197        | 7.0423  | 500  | 0.1769          | 0.9317   | 0.8268 | 0.8312    | 0.8224 |
| 0.099         | 8.4507  | 600  | 0.1841          | 0.9313   | 0.8189 | 0.8580    | 0.7833 |
| 0.0748        | 9.8592  | 700  | 0.1739          | 0.9359   | 0.8366 | 0.8457    | 0.8277 |
| 0.0706        | 11.2676 | 800  | 0.1762          | 0.9373   | 0.8399 | 0.8506    | 0.8295 |
| 0.0865        | 12.6761 | 900  | 0.1766          | 0.9408   | 0.8486 | 0.8611    | 0.8366 |
| 0.061         | 14.0845 | 1000 | 0.1852          | 0.9380   | 0.8445 | 0.8401    | 0.8490 |
| 0.0449        | 15.4930 | 1100 | 0.1795          | 0.9401   | 0.8482 | 0.8528    | 0.8437 |
| 0.0488        | 16.9014 | 1200 | 0.2065          | 0.9310   | 0.8253 | 0.8283    | 0.8224 |
| 0.0483        | 18.3099 | 1300 | 0.1977          | 0.9377   | 0.8427 | 0.8434    | 0.8419 |
| 0.0317        | 19.7183 | 1400 | 0.2006          | 0.9370   | 0.8395 | 0.8478    | 0.8313 |
| 0.0411        | 21.1268 | 1500 | 0.2068          | 0.9363   | 0.8368 | 0.8498    | 0.8242 |
| 0.0512        | 22.5352 | 1600 | 0.2056          | 0.9391   | 0.8446 | 0.8545    | 0.8348 |
| 0.0329        | 23.9437 | 1700 | 0.2127          | 0.9338   | 0.8294 | 0.8479    | 0.8117 |
| 0.0197        | 25.3521 | 1800 | 0.2122          | 0.9335   | 0.8286 | 0.8463    | 0.8117 |
| 0.0316        | 26.7606 | 1900 | 0.2050          | 0.9373   | 0.8399 | 0.8506    | 0.8295 |
| 0.0133        | 28.1690 | 2000 | 0.2019          | 0.9408   | 0.8495 | 0.8571    | 0.8419 |
| 0.0181        | 29.5775 | 2100 | 0.2031          | 0.9401   | 0.8474 | 0.8566    | 0.8384 |


### Framework versions

- Transformers 4.44.2
- Pytorch 2.5.0+cu121
- Datasets 3.1.0
- Tokenizers 0.19.1