alexandrainst/da-offensive-detection-base

Danish Offensive Text Detection based on XLM-Roberta-Base

This model is a fine-tuned version of xlm-roberta-base on a dataset consisting of approximately 5 million Facebook comments on DR's public Facebook pages. The labels have been automatically generated using weak supervision, based on the Snorkel framework.

The model achieves SOTA on a test set consisting of 600 Facebook comments annotated using majority vote by three annotators, of which 35.8% were labelled as offensive:

Using the model

You can use the model simply by running the following:

>>> from transformers import pipeline
>>> offensive_text_pipeline = pipeline(model="alexandrainst/da-offensive-detection-base")
>>> offensive_text_pipeline("Din store idiot")
[{'label': 'Offensive', 'score': 0.9997463822364807}]

Processing multiple documents at the same time can be done as follows:

>>> offensive_text_pipeline(["Din store idiot", "ej hvor godt :)"])
[{'label': 'Offensive', 'score': 0.9997463822364807}, {'label': 'Not offensive', 'score': 0.9996451139450073}]

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• —learning_rate: 2e-05
• —trainbatchsize: 32
• —evalbatchsize: 32
• —gradientaccumulationsteps: 1
• —totaltrainbatch_size: 32
• —seed: 4242
• —optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• —lrschedulertype: linear
• —max_steps: 500000
• —fp16: True
• —eval_steps: 1000
• —earlystoppingpatience: 100

Framework versions

• —Transformers 4.20.1
• —Pytorch 1.11.0+cu113
• —Datasets 2.3.2
• —Tokenizers 0.12.1