Unable to Implement Guardrails AI on GPT4All

[Abdullah7700]

Describe the bug
I'm working on a small research project to effectively implement guardrails on open source LLMs such as GPT4All and other Nomic and Leuther models. While implementing Guardrails AI on GPT4All, I am having the following problem:

To Reproduce

Steps to reproduce the behavior:

1. Image of Rail:

2. Runtime arguments:

raw_llm_output, validated_output = guard(
model.generate,
prompt_params={"question": "Give me a brief history of the second world war?"},
temperature=0
)

Expected behavior
Having implemented Guardrails AI on ChatGPT by OpenAI, I want this code to work in a similar manner in that I get a validated output by the end which has been verified by the created Guardrails spec.

Error Message:
The callable fn passed to Guard(fn, ...) failed with the following error: GPT4All.generate() got an unexpected keyword argument 'temperature'. Make sure that fn can be called as a function that takes in a single prompt string and returns a string.

My Understanding of the Problem
From what I've been able to understand, GPT4All has the keyword 'temp' for temperature e.g.

model.generate(prompt='what is the capital of France?', temp=0)

On the other hand, Guardrails AI requires the keyword temperature parameter to be given. I think that is causing a clash.

If my assessment is incorrect, what is the problem and how to solve it? If I am right, how can this be resolved?

[irgolic]

Hey, sorry for the late reply. See here for an example of how to use GPT4All with guardrails: https://github.com/guardrails-ai/guardrails/blob/gpt4all-example/docs/examples/gpt4all-guardrails.ipynb