---
library_name: transformers
tags:
- trackio
- trackio:https://huggingface.co/spaces/ciphertext/trackio
- generated_from_trainer
metrics:
- accuracy
- f1
- precision
- recall
model-index:
- name: vijil-bias-detector-v6
  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. -->

<a href="https://huggingface.co/spaces/ciphertext/trackio" target="_blank"><img src="https://raw.githubusercontent.com/gradio-app/trackio/refs/heads/main/trackio/assets/badge.png" alt="Visualize in Trackio" title="Visualize in Trackio" style="height: 40px;"/></a>
# vijil-bias-detector-v6

This model is a fine-tuned version of [](https://huggingface.co/) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 0.1858
- Accuracy: 0.9078
- F1: 0.9116
- Precision: 0.9050
- Recall: 0.9184

## 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: 1e-05
- train_batch_size: 8
- eval_batch_size: 64
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 32
- optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 0.1
- num_epochs: 10
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch  | Step | Validation Loss | Accuracy | F1     | Precision | Recall |
|:-------------:|:------:|:----:|:---------------:|:--------:|:------:|:---------:|:------:|
| 3.9400        | 0.1610 | 100  | 0.9195          | 0.6272   | 0.5429 | 0.7432    | 0.4277 |
| 2.7437        | 0.3221 | 200  | 0.6050          | 0.7303   | 0.7588 | 0.7064    | 0.8196 |
| 2.0256        | 0.4831 | 300  | 0.4529          | 0.7601   | 0.7711 | 0.7618    | 0.7807 |
| 1.6464        | 0.6441 | 400  | 0.3759          | 0.7935   | 0.7878 | 0.8417    | 0.7403 |
| 1.3442        | 0.8052 | 500  | 0.3176          | 0.8281   | 0.8346 | 0.8317    | 0.8375 |
| 1.1079        | 0.9662 | 600  | 0.2684          | 0.8551   | 0.8637 | 0.8414    | 0.8872 |
| 1.1646        | 1.1272 | 700  | 0.2793          | 0.8478   | 0.8527 | 0.8547    | 0.8507 |
| 1.0329        | 1.2882 | 800  | 0.2549          | 0.8740   | 0.8874 | 0.8259    | 0.9588 |
| 0.9363        | 1.4493 | 900  | 0.2330          | 0.8704   | 0.8786 | 0.8529    | 0.9059 |
| 0.9202        | 1.6103 | 1000 | 0.2189          | 0.8857   | 0.8904 | 0.8836    | 0.8974 |
| 0.8535        | 1.7713 | 1100 | 0.2180          | 0.8833   | 0.8839 | 0.9109    | 0.8585 |
| 0.8225        | 1.9324 | 1200 | 0.1928          | 0.8994   | 0.9048 | 0.8866    | 0.9238 |
| 0.6711        | 2.0934 | 1300 | 0.2059          | 0.8949   | 0.9015 | 0.8755    | 0.9292 |
| 0.7200        | 2.2544 | 1400 | 0.2238          | 0.8885   | 0.8992 | 0.8448    | 0.9611 |
| 0.7050        | 2.4155 | 1500 | 0.1912          | 0.9070   | 0.9115 | 0.8987    | 0.9246 |
| 0.7165        | 2.5765 | 1600 | 0.2126          | 0.8905   | 0.8987 | 0.8621    | 0.9386 |
| 0.6671        | 2.7375 | 1700 | 0.1736          | 0.9126   | 0.9152 | 0.9199    | 0.9106 |
| 0.6122        | 2.8986 | 1800 | 0.1897          | 0.9010   | 0.9015 | 0.9290    | 0.8756 |
| 0.5532        | 3.0596 | 1900 | 0.1858          | 0.9078   | 0.9129 | 0.8935    | 0.9331 |
| 0.6201        | 3.2206 | 2000 | 0.1818          | 0.9126   | 0.9176 | 0.8962    | 0.9401 |
| 0.5598        | 3.3816 | 2100 | 0.1903          | 0.9110   | 0.9160 | 0.8959    | 0.9370 |
| 0.5556        | 3.5427 | 2200 | 0.1794          | 0.9126   | 0.9155 | 0.9166    | 0.9145 |
| 0.5496        | 3.7037 | 2300 | 0.1972          | 0.9054   | 0.9124 | 0.8762    | 0.9518 |
| 0.5310        | 3.8647 | 2400 | 0.1849          | 0.9082   | 0.9142 | 0.8856    | 0.9448 |
| 0.5049        | 4.0258 | 2500 | 0.1889          | 0.9078   | 0.9110 | 0.9106    | 0.9114 |
| 0.4742        | 4.1868 | 2600 | 0.1908          | 0.9062   | 0.9114 | 0.8914    | 0.9323 |
| 0.4311        | 4.3478 | 2700 | 0.1827          | 0.9070   | 0.9119 | 0.8945    | 0.9300 |
| 0.4675        | 4.5089 | 2800 | 0.1858          | 0.9078   | 0.9116 | 0.9050    | 0.9184 |


### Framework versions

- Transformers 5.5.0
- Pytorch 2.11.0+cu130
- Datasets 4.8.4
- Tokenizers 0.22.2