Instructions to use nvidia/Nemotron-3-Content-Safety with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use nvidia/Nemotron-3-Content-Safety with Transformers:

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

pipe = pipeline("image-text-to-text", model="nvidia/Nemotron-3-Content-Safety")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
pipe(text=messages)

# Load model directly
from transformers import AutoProcessor, AutoModelForMultimodalLM

processor = AutoProcessor.from_pretrained("nvidia/Nemotron-3-Content-Safety")
model = AutoModelForMultimodalLM.from_pretrained("nvidia/Nemotron-3-Content-Safety")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
inputs = processor.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Inference
Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use nvidia/Nemotron-3-Content-Safety with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "nvidia/Nemotron-3-Content-Safety"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "nvidia/Nemotron-3-Content-Safety",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Use Docker

docker model run hf.co/nvidia/Nemotron-3-Content-Safety

SGLang

How to use nvidia/Nemotron-3-Content-Safety 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 "nvidia/Nemotron-3-Content-Safety" \
    --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": "nvidia/Nemotron-3-Content-Safety",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

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 "nvidia/Nemotron-3-Content-Safety" \
        --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": "nvidia/Nemotron-3-Content-Safety",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Docker Model Runner
How to use nvidia/Nemotron-3-Content-Safety with Docker Model Runner:
```
docker model run hf.co/nvidia/Nemotron-3-Content-Safety
```

Nemotron-3-Content-Safety / explainability.md

varunsingh

added subcards

b47f6f6 27 days ago

preview code

Raw

History Blame Contribute Delete

5.32 kB

	# Explainability Subcard


	\| Field \| Description \|
	\| ----------------------------------------------------------------------------------------------------- \| ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- \|
	\| Intended Domain \| Multimodal Content Safety \|
	\| Model Type \| Safety Classifier \|
	\| Intended Users \| AI/ML Engineers, LLM Developers, Safety Assurance Teams \|
	\| Output \| Text \|
	\| Describe how the model works: \| Type: Finetuned Transformer (Decoder-only) working as a classifier. Backbone: Google Gemma-3-4B-it Parameters: 4B (Billion) \|
	\| Name the adversely impacted groups this has been tested to deliver comparable outcomes regardless of: \| Not Applicable \|
	\| Technical Limitations: \| • The model only accepts a single text input along with an optional image. The model does not accept more than one image. \|
	\| Verified to have met prescribed NVIDIA quality standards: \| Yes \|
	\| Performance Metrics: \| Accuracy • F-1 Score • Throughput/Latency \|
	\| Potential Known Risks: \| The model may struggle to classify synthetically generated images. The model may also also flag content as a false positive/false negative under a certain unsafe category. \|
	\| Terms of Use: \| Use of the model is governed by the [OpenMDW License Agreement, version 1.1](https://raw.githubusercontent.com/OpenMDW/OpenMDW/refs/heads/main/1.1/LICENSE.OpenMDW-1.1) (OpenMDW-1.1), [Gemma Terms of Use](https://ai.google.dev/gemma/terms) and [Gemma Prohibited Use Policy](https://ai.google.dev/gemma/prohibited_use_policy). \|

	# Explainability Subcard


	\| Field \| Description \|
	\| ----------------------------------------------------------------------------------------------------- \| ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- \|
	\| Intended Domain \| Multimodal Content Safety \|
	\| Model Type \| Safety Classifier \|
	\| Intended Users \| AI/ML Engineers, LLM Developers, Safety Assurance Teams \|
	\| Output \| Text \|
	\| Describe how the model works: \| Type: Finetuned Transformer (Decoder-only) working as a classifier. Backbone: Google Gemma-3-4B-it Parameters: 4B (Billion) \|
	\| Name the adversely impacted groups this has been tested to deliver comparable outcomes regardless of: \| Not Applicable \|
	\| Technical Limitations: \| • The model only accepts a single text input along with an optional image. The model does not accept more than one image. \|
	\| Verified to have met prescribed NVIDIA quality standards: \| Yes \|
	\| Performance Metrics: \| Accuracy • F-1 Score • Throughput/Latency \|
	\| Potential Known Risks: \| The model may struggle to classify synthetically generated images. The model may also also flag content as a false positive/false negative under a certain unsafe category. \|
	\| Terms of Use: \| Use of the model is governed by the [OpenMDW License Agreement, version 1.1](https://raw.githubusercontent.com/OpenMDW/OpenMDW/refs/heads/main/1.1/LICENSE.OpenMDW-1.1) (OpenMDW-1.1), [Gemma Terms of Use](https://ai.google.dev/gemma/terms) and [Gemma Prohibited Use Policy](https://ai.google.dev/gemma/prohibited_use_policy). \|