Distilbert Nsfw Text Classifier

🚀 NSFW Text Classification Model

This model is designed to classify text into two categories: "safe" and "nsfw" (not safe for work). It is suitable for content moderation and filtering applications, effectively helping to identify inappropriate content.

🚀 Quick Start

The model can be directly used to classify text into one of the two classes. It takes a text string as input and outputs a probability distribution over the two classes. The class with the highest probability is selected as the predicted class.

Load model directly

from transformers import AutoTokenizer, AutoModelForSequenceClassification

tokenizer = AutoTokenizer.from_pretrained("eliasalbouzidi/distilbert-nsfw-text-classifier")

model = AutoModelForSequenceClassification.from_pretrained("eliasalbouzidi/distilbert-nsfw-text-classifier")

Use a pipeline

from transformers import pipeline

pipe = pipeline("text-classification", model="eliasalbouzidi/distilbert-nsfw-text-classifier")

✨ Features

• Accurate Classification: The model has achieved high scores in F1, accuracy, precision, and recall, with an F1 score of 0.974, indicating its excellent performance in text classification.
• Based on Distilbert: Built on the Distilbert-base model, it benefits from the pre - trained features of Distilbert.
• Large - scale Training: Trained on a dataset containing 190,000 labeled text samples, ensuring its generalization ability.

📦 Installation

No specific installation steps are provided in the original document.

💻 Usage Examples

Basic Usage

from transformers import AutoTokenizer, AutoModelForSequenceClassification

tokenizer = AutoTokenizer.from_pretrained("eliasalbouzidi/distilbert-nsfw-text-classifier")
model = AutoModelForSequenceClassification.from_pretrained("eliasalbouzidi/distilbert-nsfw-text-classifier")

text = "A family hiking in the mountains"
inputs = tokenizer(text, return_tensors='pt')
outputs = model(**inputs)

Advanced Usage

from transformers import pipeline

pipe = pipeline("text-classification", model="eliasalbouzidi/distilbert-nsfw-text-classifier")
texts = ["A family hiking in the mountains", "A couple kissing passionately in bed"]
results = pipe(texts)

📚 Documentation

Model Description

The model is based on the Distilbert-base model and is trained to classify text into "safe" and "nsfw" categories. It takes text as input and outputs the probability of each class.

Uses

The model can be integrated into larger systems for content moderation or filtering, such as social media platforms, online forums, etc.

Training Data

The training data for finetuning the text classification model consists of a large corpus of text labeled with one of the two classes: "safe" and "nsfw". The dataset contains a total of 190,000 examples, with 117,000 labeled as "safe" and 63,000 labeled as "nsfw". It was assembled by scraping data from the web and utilizing existing open - source datasets. A significant portion of the dataset consists of descriptions for images and scenes. You can access the dataset: https://huggingface.co/datasets/eliasalbouzidi/NSFW-Safe-Dataset

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 1e - 05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon = 1e - 08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 600
• num_epochs: 3
• mixed_precision_training: Native AMP

Training results

We selected the checkpoint with the highest F - beta1.6 score.

Framework versions

• Transformers 4.40.1
• Pytorch 2.3.0+cu121
• Datasets 2.19.0
• Tokenizers 0.19.1

Out - of - Scope Use

It should not be used for any illegal activities.

🔧 Technical Details

The model is based on the Distilbert - base model. During training, a large - scale dataset containing 190,000 labeled text samples was used. The hyperparameters were carefully tuned to achieve high performance in F1, accuracy, precision, and recall.

📄 License

The model is licensed under the apache - 2.0 license.

Bias, Risks, and Limitations

The model may exhibit biases based on the training data used. It may not perform well on text that is written in languages other than English. It may also struggle with sarcasm, irony, or other forms of figurative language. The model may produce false positives or false negatives, which could lead to incorrect categorization of text.

Recommendations

⚠️ Important Note

The model has limitations and biases. Users should be aware of these and use it accordingly.

💡 Usage Tip

Users should be prepared to handle false positives and false negatives. It is recommended to fine - tune the model for specific downstream tasks and to evaluate its performance on relevant datasets.

Contact

Please reach out to eliasalbouzidi@gmail.com if you have any questions or feedback.