Instructions to use Ursulalala/HomeGuard-8B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Ursulalala/HomeGuard-8B with Transformers:

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

pipe = pipeline("image-text-to-text", model="Ursulalala/HomeGuard-8B")
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("Ursulalala/HomeGuard-8B")
model = AutoModelForMultimodalLM.from_pretrained("Ursulalala/HomeGuard-8B", device_map="auto")
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]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use Ursulalala/HomeGuard-8B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Ursulalala/HomeGuard-8B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Ursulalala/HomeGuard-8B",
		"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/Ursulalala/HomeGuard-8B

SGLang

How to use Ursulalala/HomeGuard-8B 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 "Ursulalala/HomeGuard-8B" \
    --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": "Ursulalala/HomeGuard-8B",
		"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 "Ursulalala/HomeGuard-8B" \
        --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": "Ursulalala/HomeGuard-8B",
		"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 Ursulalala/HomeGuard-8B with Docker Model Runner:
```
docker model run hf.co/Ursulalala/HomeGuard-8B
```

Ursulalala commited on Mar 23

Commit

0766d4c

verified ·

1 Parent(s): 7faf33a

Update README.md

Browse files

Files changed (1) hide show

README.md +2 -28

README.md CHANGED Viewed

@@ -22,11 +22,7 @@ This checkpoint corresponds to the **8B step-RFT model** used in the HomeGuard f
 ## Model Summary
-HomeGuard focuses on scenarios where a seemingly benign instruction becomes unsafe because of object attributes, spatial relations, or latent environmental conditions. Typical examples include:
-- heating food when metal is inside the microwave
-- using electrical devices near water
-- executing routine motions in cluttered environments with tripping hazards
 Compared with generic VLMs, HomeGuard is specialized for:
@@ -41,24 +37,8 @@ This model is derived from **Qwen3-VL-8B-Thinking** and trained within the HomeG
 Training setup summarized from the released training configuration:
 - Base model: `Qwen/Qwen3-VL-8B-Thinking`
-- Training stage: step-level RFT / GRPO-style optimization in the HomeGuard pipeline
 - Training data: HomeSafe
-- Epochs: `1`
-- Per-device batch size: `1`
-- Gradient accumulation steps: `2`
-- Learning rate: `1e-6`
-- Warmup ratio: `0.1`
-- Save step for this checkpoint: `500`
-- Precision: `bf16`
-Reward weights for this checkpoint:
-- safe accuracy: `1.0`
-- safety hazard match: `0.5`
-- principle accuracy: `2.0`
-- IoU target object: `2.0`
-- IoU constraint object: `2.0`
-- format reward: `1.0`
 ## Intended Use
@@ -94,12 +74,6 @@ For full prompting, evaluation, and application examples, please refer to the Ho
 - Dataset: [Ursulalala/HomeSafe](https://huggingface.co/datasets/Ursulalala/HomeSafe)
 - Base model: [Qwen/Qwen3-VL-8B-Thinking](https://huggingface.co/Qwen/Qwen3-VL-8B-Thinking)
-## Limitations
-- This model is specialized for household contextual safety and should not be treated as a complete safety solution for all robotics settings.
-- Real-world deployment still requires additional system-level safeguards, environment checks, and action constraints.
-- Performance may vary across sensors, domains, and task distributions outside HomeSafe.
 ## Citation
 If you use this model, please cite the HomeGuard paper:

 ## Model Summary
+HomeGuard focuses on scenarios where a seemingly benign instruction becomes unsafe because of object attributes, spatial relations, or latent environmental conditions.
 Compared with generic VLMs, HomeGuard is specialized for:
 Training setup summarized from the released training configuration:
 - Base model: `Qwen/Qwen3-VL-8B-Thinking`
+- Training stage: step-level RFT + GRPO-style optimization in the HomeGuard pipeline
 - Training data: HomeSafe
 ## Intended Use
 - Dataset: [Ursulalala/HomeSafe](https://huggingface.co/datasets/Ursulalala/HomeSafe)
 - Base model: [Qwen/Qwen3-VL-8B-Thinking](https://huggingface.co/Qwen/Qwen3-VL-8B-Thinking)
 ## Citation
 If you use this model, please cite the HomeGuard paper: