---
library_name: transformers
license: apache-2.0
base_model: facebook/dinov2-small
tags:
- generated_from_trainer
metrics:
- accuracy
- f1
- recall
- precision
model-index:
- name: dinov2_Liveness_detection_v2.2
  results: []
---

<!-- 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. -->

[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="200" height="32"/>](https://wandb.ai/nguyenkhoaht002/liveness_detection/runs/b39fcrkm)
# dinov2_Liveness_detection_v2.2

This model is a fine-tuned version of [facebook/dinov2-small](https://huggingface.co/facebook/dinov2-small) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 0.1307
- Accuracy: 0.9781
- F1: 0.9781
- Recall: 0.9781
- Precision: 0.9783

## 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: 128
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 5
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch  | Step | Validation Loss | Accuracy | F1     | Recall | Precision |
|:-------------:|:------:|:----:|:---------------:|:--------:|:------:|:------:|:---------:|
| 0.3279        | 0.2048 | 128  | 0.2858          | 0.8749   | 0.8772 | 0.8749 | 0.8773    |
| 0.2389        | 0.4096 | 256  | 0.2696          | 0.8881   | 0.8819 | 0.8881 | 0.9196    |
| 0.186         | 0.6144 | 384  | 0.1614          | 0.9383   | 0.9386 | 0.9383 | 0.9381    |
| 0.2048        | 0.8192 | 512  | 0.1568          | 0.9404   | 0.9411 | 0.9404 | 0.9415    |
| 0.1662        | 1.024  | 640  | 0.1474          | 0.9426   | 0.9433 | 0.9426 | 0.9436    |
| 0.1257        | 1.2288 | 768  | 0.1186          | 0.9578   | 0.9573 | 0.9578 | 0.9604    |
| 0.1215        | 1.4336 | 896  | 0.1202          | 0.9556   | 0.9560 | 0.9556 | 0.9561    |
| 0.0917        | 1.6384 | 1024 | 0.1045          | 0.9611   | 0.9611 | 0.9611 | 0.9611    |
| 0.1256        | 1.8432 | 1152 | 0.0971          | 0.9633   | 0.9630 | 0.9633 | 0.9645    |
| 0.0676        | 2.048  | 1280 | 0.1524          | 0.9487   | 0.9477 | 0.9487 | 0.9545    |
| 0.0458        | 2.2528 | 1408 | 0.1149          | 0.9641   | 0.9643 | 0.9641 | 0.9642    |
| 0.0462        | 2.4576 | 1536 | 0.1233          | 0.9630   | 0.9632 | 0.9630 | 0.9631    |
| 0.0453        | 2.6624 | 1664 | 0.1030          | 0.9671   | 0.9670 | 0.9671 | 0.9679    |
| 0.0631        | 2.8672 | 1792 | 0.0896          | 0.967    | 0.9672 | 0.967  | 0.9671    |
| 0.0358        | 3.072  | 1920 | 0.0966          | 0.9735   | 0.9734 | 0.9735 | 0.9738    |
| 0.0229        | 3.2768 | 2048 | 0.1250          | 0.9675   | 0.9676 | 0.9675 | 0.9676    |
| 0.0272        | 3.4816 | 2176 | 0.1148          | 0.9691   | 0.9693 | 0.9691 | 0.9692    |
| 0.0253        | 3.6864 | 2304 | 0.1130          | 0.9757   | 0.9755 | 0.9757 | 0.9761    |
| 0.0249        | 3.8912 | 2432 | 0.1091          | 0.9716   | 0.9717 | 0.9716 | 0.9715    |
| 0.0049        | 4.096  | 2560 | 0.1420          | 0.9756   | 0.9756 | 0.9756 | 0.9755    |
| 0.0159        | 4.3008 | 2688 | 0.1423          | 0.9775   | 0.9774 | 0.9775 | 0.9777    |
| 0.0026        | 4.5056 | 2816 | 0.1454          | 0.9774   | 0.9773 | 0.9774 | 0.9776    |
| 0.0059        | 4.7104 | 2944 | 0.1445          | 0.9785   | 0.9785 | 0.9785 | 0.9785    |
| 0.0011        | 4.9152 | 3072 | 0.1307          | 0.9781   | 0.9781 | 0.9781 | 0.9783    |

### Evaluate results

- Accuaracy: 0.81
- F1: 0.86
- Recall: 0.85
- Precision: 0.65

- APCER: 0.2001
- BPCER: 0.1458
- ACER: 0.1729

### Framework versions

- Transformers 4.44.2
- Pytorch 2.4.1+cu121
- Datasets 3.2.0
- Tokenizers 0.19.1