Instructions to use prem-research/MiniGuard-v0.1 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use prem-research/MiniGuard-v0.1 with Transformers:

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

pipe = pipeline("text-generation", model="prem-research/MiniGuard-v0.1")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("prem-research/MiniGuard-v0.1")
model = AutoModelForCausalLM.from_pretrained("prem-research/MiniGuard-v0.1", device_map="auto")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.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(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use prem-research/MiniGuard-v0.1 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "prem-research/MiniGuard-v0.1"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "prem-research/MiniGuard-v0.1",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/prem-research/MiniGuard-v0.1

SGLang

How to use prem-research/MiniGuard-v0.1 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 "prem-research/MiniGuard-v0.1" \
    --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": "prem-research/MiniGuard-v0.1",
		"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 "prem-research/MiniGuard-v0.1" \
        --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": "prem-research/MiniGuard-v0.1",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use prem-research/MiniGuard-v0.1 with Docker Model Runner:
```
docker model run hf.co/prem-research/MiniGuard-v0.1
```

ojus1 commited on Dec 15, 2025

Commit

f341bc8

verified ·

1 Parent(s): 78278c4

Update README.md

Browse files

Files changed (1) hide show

README.md +13 -7

README.md CHANGED Viewed

@@ -17,7 +17,7 @@ tags:
 MiniGuard-v0.1 is a compact content safety classifier fine-tuned from [Qwen3-0.6B](https://huggingface.co/Qwen/Qwen3-0.6B). It classifies content in both, User inputs (prompt classification) and LLM responses (response classification), outputting whether a given prompt or response is safe or unsafe, along with the violated safety categories.
-MiniGuard-v0.1 achieves **~99% of Nemotron-Guard-8B's benchmark accuracy** with **13x fewer parameters**.
 ## Compatibility
@@ -56,14 +56,16 @@ Dataset - English subset test split of [nvidia/Nemotron-Safety-Guard-Dataset-v3]
 Evaluated on out-of-distribution production data containing real user queries. Cost estimated based on H200 GPU pricing ($7.91/hour) at maximum concurrency with P95 latency SLA of <500ms.
-| Metric | MiniGuard-v0.1 | Nemotron-Guard-8B-v3 |
-|--------|----------------|----------------------|
-| Relative Macro F1 | 91.1% | 100% |
-| Cost per 1M requests | **$15.54** | $46.93 |
-| Cost Savings | **67%** | - |
-MiniGuard-v0.1 achieves 91.1% relative performance on out-of-distribution data while costing **67% less** to serve.
 ### Ablation Study
@@ -77,6 +79,10 @@ Impact of techniques on out-of-distribution production data (Relative Macro F1 c
 | + Targeted Synthetic Data | 0.6B | 87.2% | +1.6% |
 | + Soup (top-3) [MiniGuard-v0.1] | 0.6B | 92.3% | +5.1% |
 | + FP8 | 0.6B | 91.1% | -1.2% |
 | Nemotron-Guard-8B-v3 | 8B | 100% | reference |
 #### In-Distribution

 MiniGuard-v0.1 is a compact content safety classifier fine-tuned from [Qwen3-0.6B](https://huggingface.co/Qwen/Qwen3-0.6B). It classifies content in both, User inputs (prompt classification) and LLM responses (response classification), outputting whether a given prompt or response is safe or unsafe, along with the violated safety categories.
+MiniGuard-v0.1 achieves **~99% of Nemotron-Guard-8B's benchmark accuracy** with **13x fewer parameters** and **outperforms Qwen3Guard-8B** (a specialized 8B safety model) **by 9 percentage points** on production data.
 ## Compatibility
 Evaluated on out-of-distribution production data containing real user queries. Cost estimated based on H200 GPU pricing ($7.91/hour) at maximum concurrency with P95 latency SLA of <500ms.
+| Model | Parameters | Rel. Macro F1 | Cost per 1M requests | Cost Savings |
+|--------|------------|---------------|----------------------|--------------|
+| **MiniGuard-v0.1** | **0.6B** | **91.1%** | **$15.54** | **67%** |
+| Qwen3Guard-Gen-0.6B | 0.6B | 72.1% | - | - |
+| Qwen3Guard-Gen-4B | 4B | 78.0% | - | - |
+| Qwen3Guard-Gen-8B | 8B | 82.1% | - | - |
+| Nemotron-Guard-8B-v3 | 8B | 100% | $46.93 | baseline |
+MiniGuard-v0.1 achieves 91.1% relative performance on out-of-distribution data while costing **67% less** to serve. **Notably, our 0.6B fine-tuned model outperforms all Qwen3Guard models by significant margins**, including the 8B version (82.1%), demonstrating that targeted fine-tuning is more effective than simply using larger pretrained safety models.
 ### Ablation Study
 | + Targeted Synthetic Data | 0.6B | 87.2% | +1.6% |
 | + Soup (top-3) [MiniGuard-v0.1] | 0.6B | 92.3% | +5.1% |
 | + FP8 | 0.6B | 91.1% | -1.2% |
+| **Comparison Baselines:** |
+| Qwen3Guard-Gen-0.6B | 0.6B | 72.1% | - |
+| Qwen3Guard-Gen-4B | 4B | 78.0% | - |
+| Qwen3Guard-Gen-8B | 8B | 82.1% | - |
 | Nemotron-Guard-8B-v3 | 8B | 100% | reference |
 #### In-Distribution