Instructions to use FrontiersMind/Nandi-Mini-600M-GuardRails with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use FrontiersMind/Nandi-Mini-600M-GuardRails with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="FrontiersMind/Nandi-Mini-600M-GuardRails", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("FrontiersMind/Nandi-Mini-600M-GuardRails", trust_remote_code=True, dtype="auto")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use FrontiersMind/Nandi-Mini-600M-GuardRails with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "FrontiersMind/Nandi-Mini-600M-GuardRails"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "FrontiersMind/Nandi-Mini-600M-GuardRails",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/FrontiersMind/Nandi-Mini-600M-GuardRails

SGLang

How to use FrontiersMind/Nandi-Mini-600M-GuardRails with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "FrontiersMind/Nandi-Mini-600M-GuardRails" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "FrontiersMind/Nandi-Mini-600M-GuardRails",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "FrontiersMind/Nandi-Mini-600M-GuardRails" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "FrontiersMind/Nandi-Mini-600M-GuardRails",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use FrontiersMind/Nandi-Mini-600M-GuardRails with Docker Model Runner:
```
docker model run hf.co/FrontiersMind/Nandi-Mini-600M-GuardRails
```

Nandi-Mini-150M-GuardRails

Introduction

Nandi-Mini-600M-GuardRails is a lightweight multilingual (English & Hindi) safety classification model that detects unsafe or policy-violating content in user prompts and AI responses across multiple harm categories.

🚀 Usage

!pip install transformers=='5.4.0'

from transformers import AutoTokenizer, AutoModelForCausalLM
import torch
import json

model_name = "FrontiersMind/Nandi-Mini-600M-GuardRails"

device = "cuda" if torch.cuda.is_available() else "cpu"

tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained(
    model_name,
    trust_remote_code=True,
    dtype=torch.bfloat16
).to(device).eval()


def classify_safety(prompt, response=None):

    content = {"prompt": prompt}

    if response is not None:
        content["response"] = response

    messages = [
        {
            "role": "user",
            "content": content
        }
    ]

    prompt = tokenizer.apply_chat_template(
        messages,
        tokenize=False,
        add_generation_prompt=True
    )

    inputs = tokenizer(prompt, return_tensors="pt").to(model.device)

    generated_ids = model.generate(
        **inputs,
        max_new_tokens=200,
        do_sample=False,
        temperature=0.0,
    )

    generated_ids = [
        output_ids[len(input_ids):]
        for input_ids, output_ids in zip(inputs.input_ids, generated_ids)
    ]

    response = tokenizer.batch_decode(
        generated_ids,
        skip_special_tokens=True
    )[0]

    return json.loads(response)


result = classify_safety(
    prompt="Tell me how to kill someone.",
)

print(result)