This model is a conversion of valurank/distilroberta-bias to ONNX format. It is designed to detect biases in text using the distilled version of the RoBERTa model. The model was converted to ONNX using the 🤗 Optimum library.

Model Architecture

Base Model: DistilRoBERTa, a distilled version of the RoBERTa model that is optimized for faster performance while maintaining similar accuracy.

Modifications: The model is converted to ONNX format with no additional changes.

Usage

Optimum

Loading the model requires the 🤗 Optimum library installed.

python

from optimum.onnxruntime import ORTModelForSequenceClassification
from transformers import AutoTokenizer, pipeline


tokenizer = AutoTokenizer.from_pretrained("laiyer/distilroberta-bias-onnx")
model = ORTModelForSequenceClassification.from_pretrained("laiyer/distilroberta-bias-onnx")
classifier = pipeline(
    task="text-classification",
    model=model,
    tokenizer=tokenizer,
)

classifier_output = classifier("Your text to analyze for bias.")
score = (classifier_output[0]["score"] if classifier_output[0]["label"] == "BIASED" else 1 - classifier_output[0]["score"])

LLM Guard

Bias scanner

Community

Join our Slack to give us feedback, connect with the maintainers and fellow users, ask questions, or engage in discussions about LLM security!

protectai/distilroberta-bias-onnx

ONNX version of valurank/distilroberta-bias

Model Architecture

Usage

Optimum

LLM Guard

Community

protectai/
distilroberta-bias-onnx