Uzyau/distilbert-toxicity-onnx

DistilBERT multilingual toxicity classifier (ONNX)

ONNX-converted version of citizenlab/distilbert-base-multilingual-cased-toxicity for use with InferenceWall.

Usage with InferenceWall

pip install inferwall[standard]
inferwall models install --profile standard

Usage with ONNX Runtime

from optimum.onnxruntime import ORTModelForSequenceClassification
from transformers import AutoTokenizer

model = ORTModelForSequenceClassification.from_pretrained("inferwall/distilbert-toxicity-onnx")
tokenizer = AutoTokenizer.from_pretrained("inferwall/distilbert-toxicity-onnx")

inputs = tokenizer("your text here", return_tensors="pt")
outputs = model(**inputs)

Labels

not_toxic, toxic

Conversion

Converted from PyTorch using Hugging Face Optimum:

optimum-cli export onnx --model citizenlab/distilbert-base-multilingual-cased-toxicity --task text-classification output/

Original Model

• Source: citizenlab/distilbert-base-multilingual-cased-toxicity
• Converted from PyTorch. Original by Citizen Lab.