Text Classification
Transformers
ONNX
Safetensors
Transformers.js
modernbert
abacusai
toxicity
content-moderation
trust-and-safety
perspective-api
hate-speech
multi-label-classification
text-embeddings-inference
Instructions to use abacusai/Toxicity-Lens with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use abacusai/Toxicity-Lens with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-classification", model="abacusai/Toxicity-Lens")# pip install -U transformers accelerate # Load model directly from transformers import AutoTokenizer, AutoModelForSequenceClassification tokenizer = AutoTokenizer.from_pretrained("abacusai/Toxicity-Lens") model = AutoModelForSequenceClassification.from_pretrained("abacusai/Toxicity-Lens", device_map="auto") - Transformers.js
How to use abacusai/Toxicity-Lens with Transformers.js:
// npm i @huggingface/transformers import { pipeline } from '@huggingface/transformers'; // Allocate pipeline const pipe = await pipeline('text-classification', 'abacusai/Toxicity-Lens'); - Notebooks
- Google Colab
- Kaggle
Download NOTICE from abacusai/Toxicity-Lens: direct link, hf CLI and curl.
- Browser
- Download file 3.96 kB
-
https://huggingface.co/abacusai/Toxicity-Lens/resolve/main/NOTICE
- Command line
-
hf download hf://abacusai/Toxicity-Lens/NOTICE
-
curl -L -o NOTICE https://huggingface.co/abacusai/Toxicity-Lens/resolve/main/NOTICE
3.96 kB
| Toxicity-Lens | |
| Copyright 2026 Abacus.AI | |
| This model is a finetune of jhu-clsp/mmBERT-base, trained with the data listed in README.md section 7. | |
| Perspective is a trademark of Google LLC. This project is not affiliated with or endorsed by Google or Jigsaw. | |
| ------------------------------------------------------------------------ | |
| jhu-clsp/mmBERT-base (Marone et al., 2025): MIT License, as declared on its model card (the repository publishes no copyright line). | |
| Its tokenizer is Gemma 2's; see the Gemma Terms of Use, https://ai.google.dev/gemma/terms. | |
| Permission is hereby granted, free of charge, to any person obtaining a copy | |
| of this software and associated documentation files (the "Software"), to deal | |
| in the Software without restriction, including without limitation the rights | |
| to use, copy, modify, merge, publish, distribute, sublicense, and/or sell | |
| copies of the Software, and to permit persons to whom the Software is | |
| furnished to do so, subject to the following conditions: | |
| The above copyright notice and this permission notice shall be included in all | |
| copies or substantial portions of the Software. | |
| THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR | |
| IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, | |
| FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE | |
| AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER | |
| LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, | |
| OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE | |
| SOFTWARE. | |
| ------------------------------------------------------------------------ | |
| MLMA hate speech dataset (https://github.com/HKUST-KnowComp/MLMA_hate_speech): MIT License | |
| Copyright (c) 2019 HKUST-KnowComp | |
| Permission is hereby granted, free of charge, to any person obtaining a copy | |
| of this software and associated documentation files (the "Software"), to deal | |
| in the Software without restriction, including without limitation the rights | |
| to use, copy, modify, merge, publish, distribute, sublicense, and/or sell | |
| copies of the Software, and to permit persons to whom the Software is | |
| furnished to do so, subject to the following conditions: | |
| The above copyright notice and this permission notice shall be included in all | |
| copies or substantial portions of the Software. | |
| THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR | |
| IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, | |
| FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE | |
| AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER | |
| LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, | |
| OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE | |
| SOFTWARE. | |
| ------------------------------------------------------------------------ | |
| This model contains information from FineWeb-2 (https://huggingface.co/datasets/HuggingFaceFW/fineweb-2), which is made available under the ODC Attribution License (https://opendatacommons.org/licenses/by/1-0/), and from HPLT 2.0 (https://huggingface.co/datasets/HPLT/HPLT2.0_cleaned), whose packaging is CC0 1.0. | |
| AfriHate (https://huggingface.co/datasets/afrihate/afrihate), Muhammad et al., NAACL 2025: Apache 2.0 on Hugging Face, CC BY on GitHub. The authors ask that models built with it not be used for automated removal. | |
| Civil Comments (google/civil_comments) and Horizon-Labs/civil-comments-multilingual: CC0 1.0. Qwen/Qwen3.8-27B-FP8, used to translate and label training data: Apache 2.0. | |
| RealToxicityPrompts (https://huggingface.co/datasets/allenai/real-toxicity-prompts), Gehman et al., 2020: Apache 2.0. PolygloToxicityPrompts (https://huggingface.co/datasets/ToxicityPrompts/PolygloToxicityPrompts), Jain et al., 2024: CC BY 4.0 per its card metadata; its card text names the AI2 ImpACT License - Low Risk Artifacts. Both carry scores from the Perspective API (Google LLC). | |