README.md

---
license: apache-2.0
base_model: bert-large-uncased
tags:
- adult text classification
- adult
- adult-content
metrics:
- accuracy
model-index:
- name: bert-large-uncased-Adult-Text-Classifier
  results: []
datasets:
- valurank/Adult-content-dataset
language:
- en
pipeline_tag: text-classification
---

# bert-large-uncased-Adult-Text-Classifier
This model is a fine-tuned version of [bert-large-uncased](https://huggingface.co/bert-large-uncased) on the [valurank/Adult-content-dataset](https://huggingface.co/datasets/valurank/Adult-content-dataset). It has been trained to classify text into categories related to adult content.
It achieves the following results on the evaluation set:
- Loss: 0.1257
- Accuracy: 0.9824

## Model description

The model is based on BERT (Bidirectional Encoder Representations from Transformers), specifically the uncased version which does not differentiate between capital and lowercase letters. It has been fine-tuned using the Adult Content Dataset to classify text accurately.

## Intended uses & limitations

This model can be used for various applications where identifying adult content in text is necessary, such as content filtering, moderation systems, or parental controls. However, it's essential to note that no model is perfect, and this model may still make errors in classification. Additionally, the model's performance may vary depending on the context and language used in the text.

## Training and evaluation data

The model has been trained on the Valurank Adult Content Dataset, which contains a labeled collection of text data categorized into adult and non-adult content. It was trained using 80% of data for training and rest for validation.

## Training procedure

### Training hyperparameters

The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 10

### Training results

| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|:-------------:|:-----:|:----:|:---------------:|:--------:|
| No log        | 1.0   | 43   | 0.1197          | 0.9588   |
| No log        | 2.0   | 86   | 0.1943          | 0.9529   |
| No log        | 3.0   | 129  | 0.0942          | 0.9765   |
| No log        | 4.0   | 172  | 0.1308          | 0.9765   |
| No log        | 5.0   | 215  | 0.1178          | 0.9765   |
| No log        | 6.0   | 258  | 0.1159          | 0.9824   |
| No log        | 7.0   | 301  | 0.1175          | 0.9824   |
| No log        | 8.0   | 344  | 0.1209          | 0.9824   |
| No log        | 9.0   | 387  | 0.1243          | 0.9824   |
| No log        | 10.0  | 430  | 0.1257          | 0.9824   |


### Framework versions

- Transformers 4.35.2
- Pytorch 2.1.0+cu121
- Datasets 2.16.1
- Tokenizers 0.15.0

This model card provides an overview of the model's architecture, training procedure, and performance metrics. It serves as a reference for users interested in utilizing or further understanding the capabilities and limitations of the bert-large-uncased-Adult-Text-Classifier model.