oxyapi/albert-moderation-001

Introduction

Albert Moderation 001 is a fine-tuned version of the distilbert/distilbert-base-uncased a distilled version of BERT, smaller and faster.

Developed by Oxygen (oxyapi), with contributions from TornadoSoftwares, Albert Moderation 001 allows you to moderate text content very quickly and efficiently across multiple categories

Model Details

• Model Name: Albert Moderation 001
• Model ID: oxyapi/albert-moderation-001
• Base Model: distilbert/distilbert-base-uncased
• Model Type: Text classification, Moderation
• License: Apache-2.0
• Language: English

Features

• Categories: This model classifies text data into 11 different categories: harassment, harassment/threat, sexual, hate, hate/threat, self-harm/intent, self-harm/instructions, self-harm, sexual/minors, violence, violence/graphic
• Efficient: Compact model size allows for faster inference and reduced computational resources.

Metadata

• Owned by: Oxygen (oxyapi)
• Contributors: TornadoSoftwares
• Description: A fast and lightweight moderation model based on BERT

Usage

To utilize Albert Moderation 001 for text classification, you can load the model using the Hugging Face Transformers library:

from transformers import pipeline
text = "Hey little shit, GIVE ME YOUR SNACK !"
classifier = pipeline("text-classification", model="oxyapi/albert-moderation-001", tokenizer="oxyapi/albert-moderation-001")
result = classifier(text,top_k=len(classifier.model.config.id2label))
print(result)

License

This model is licensed under the Apache 2.0 License.

Citation

If you find Albert Moderation 001 useful in your research or applications, please cite it as:

@misc{albertmoderation0012025,
  title={Albert Moderation 001: A fast and lightweight moderation model based on BERT},
  author={Oxygen (oxyapi)},
  year={2024},
  howpublished={\url{https://huggingface.co/oxyapi/albert-moderation-001}},
}