metadata
license: apache-2.0
widget:
- text: These are nice flowers
- text: What the hell
- text: You really suck, dude
- text: How to put screw thread in furniture?
- text: >-
The vacuum cleaner began to suck up the dust from the carpet, making the
room much cleaner.
metrics:
- name: Accuracy
type: accuracy
value: 0.9748
- name: Precision
type: precision
value: 0.9331
- name: Recall
type: recall
value: 0.9416
- name: F1 Score
type: f1
value: 0.9373
- name: AUC-ROC
type: roc_auc
value: 0.9955
base_model: distilbert/distilbert-base-uncased
datasets:
- tarekziade/profanity
library_name: transformers
Fine-tuned model that detects profanity in text.
Inspired from https://victorzhou.com/blog/better-profanity-detection-with-scikit-learn/
The model was trained with the dataset from that project.
Usage example with Python:
from transformers import pipeline
classifier = pipeline("sentiment-analysis", model="tarekziade/pardonmyai")
print(classifier("These are beautiful flowers"))
Usage example with Transformers.js:
import { pipeline } from '@xenova/transformers';
let pipe = await pipeline('sentiment-analysis', model='tarekziade/pardonmyai');
let out = await pipe('These are beautiful flowers');
Source code and data: https://github.com/tarekziade/pardonmyai
metrics:
- Accuracy: 0.9748
- Precision: 0.9331
- Recall: 0.9416
- F1 Score: 0.9373
- AUC-ROC: 0.9955
There's a tiny version available: https://huggingface.co/tarekziade/pardonmyai-tiny