NeMo-Guardrails
NeMo-Guardrails copied to clipboard
NVIDIA
NeMo Guardrails is an open-source toolkit for easily adding programmable guardrails to LLM-based conversational systems.
Hi I was running actions.py via self_fact_check but unsuccessful. As I am getting below error. Error 'LLMTaskManager' object has no attribute 'get_stop_tokens' while execution self_check_facts Traceback (most recent call last):...
Hi @drazvan , Its having another issue on 0.8.0 tag. When I run llama 7b as nemoguardrails chat --config ~/venv/lib/python3.10/site-packages/examples/configs/llm/hf_pipeline_llama2/ (I changed occurance of 13b to 7b in the config.yml...
Hello, I am facing an error when initializing LLMRails. Here is the error: `srun: error: task 0: Illegal instruction (core dumped)`. I think the error is related to `asyncio`. After...
Error while following the [example](https://github.com/NVIDIA/NeMo-Guardrails/blob/develop/docs/user_guides/langchain/chain-with-guardrails/chain-with-guardrails.ipynb) in the repo 
I want to perform a preprocessing step in some flows. How can I modify the user input message across these flows? Specifically, I aim to alter the last_user_message and proceed...
Hi! I would like to use embeddings_only dialog rails which will fall into fallback: ``` define flow ... priority 0.1 user ... bot responds truthfully ```` when semantic similarity between...
Failed to use Nemo guardrail with other languages, the responses are always like `I don't understand `, so it does recognize the language of input, but it still provides no...
Hi, I want to use the 'nemoguardrails server' in the same way I use 'nemoguardrails chat' But, when I set an 'init' function inside my config.py, the passed parameter to...
The problem appears when attempting to integrate some models (`llama2-7b`, `falcon-7b`, `vicuna-13b`, `llama2-13b`, `mistral-7b`) with some of the moderation tasks such as input validation, fact-checking or jailbreak detection. The official...
If the output of a chain is a string, and it if the guardrails layer is blocking the input, or the output, the response format becomes a dict with the...
