Skip to main content
Do you like Artifex? Give it a ⭐ star on GitHub!

Perform Inference with Guardrail

Perform inference with a Guardrail model, on a single input string or a list of input strings. For each input string, this will return one of two predicted labels: safe if the chatbot is allowed to generate the input string or unsafe if the chatbot is not allowed to generate the input string.

Arguments

  • text
    str | list[str]

    A string or a list of strings to classify. The model will return a label for each input string.

Response

A list[str], each being either "safe" or "unsafe", which indicate whether the chatbot is allowed to generate the input string or not.

from artifex import Artifex

guardrail = Artifex().guardrail

# NOTE: train your Guardrail model before 
# using it for inference.

label = guardrail("This is a sample safe sentence.")
print(label)
["safe"]