Instructions to use mlninad/PII-Shield with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use mlninad/PII-Shield with Transformers:

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

pipe = pipeline("text-generation", model="mlninad/PII-Shield")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("mlninad/PII-Shield")
model = AutoModelForCausalLM.from_pretrained("mlninad/PII-Shield", 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 mlninad/PII-Shield with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/mlninad/PII-Shield

SGLang

How to use mlninad/PII-Shield 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 "mlninad/PII-Shield" \
    --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": "mlninad/PII-Shield",
		"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 "mlninad/PII-Shield" \
        --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": "mlninad/PII-Shield",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use mlninad/PII-Shield with Docker Model Runner:
```
docker model run hf.co/mlninad/PII-Shield
```

PII-Shield / README.md

mlninad

Update README.md

1aa529e verified over 1 year ago

preview code

Raw

History Blame Contribute Delete

3.46 kB

metadata

tags:
  - text-generation-inference
  - transformers
  - llama
  - trl
license: apache-2.0
language:
  - en
base_model:
  - meta-llama/Llama-3.2-3B-Instruct

🛡️ PII-Shield

Your Intelligent Guardian for Personal Data Protection

🌟 What is PII-Shield?

PII-Shield is your cutting-edge solution for protecting sensitive information in text data. Powered by advanced transformer architecture, it's your first line of defense against unintended PII exposure.

🎯 Core Capabilities

🔍 Smart Detection

"Regular text with sarah.smith@email.com" → "Regular text with [EMAIL_1]"

🎭 Intelligent Masking

"Call John at (555) 123-4567" → "Call [PERSON_1] at [PHONE_1]"

📊 Structured Mapping

Original → Masked → JSON Mapping

🚀 Model Architecture

🧠 Two-Stage Intelligence

⚡ Supported PII Categories

Category	Icon	Example
Names	👤	John Smith
Emails	📧	user@domain.com
Phones	📱	(555) 123-4567
Addresses	🏠	123 Privacy St
SSN	🔢	XXX-XX-XXXX
Credit Cards	💳	XXXX-XXXX-XXXX
DOB	📅	MM/DD/YYYY
IPs	🌐	192.168.1.1

💫 How It Works

🎯 Detection Phase

def detect_pii(text: str) -> List[Entity]:
    """
    🔍 Intelligent PII detection
    Returns list of identified entities
    """
    pass

🎭 Masking Phase

def mask_pii(text: str, entities: List[Entity]) -> Dict:
    """
    🛡️ Smart PII masking
    Returns masked text and mapping
    """
    pass

🎮 Input/Output

📥 Input Format

{
    "text": "Your sensitive text here",
    "options": {
        "mask_format": "[TYPE_INDEX]",
        "return_mapping": true
    }
}

📤 Output Format

{
    "masked_text": "Your [TYPE_1] text here",
    "pii_mapping": [
        {
            "label": "TYPE",
            "value": "sensitive",
            "index": 1
        }
    ]
}

🚦 Performance Stats

Metric	Score	Trend
Precision	98.5%	⬆️
Recall	97.8%	⬆️
Speed	2ms/req	⬇️
Accuracy	99.1%	➡️

🛠️ Technical Requirements

🖥️ CUDA-capable GPU
💾 8GB+ VRAM
🐍 Python 3.8+
🔧 PyTorch 2.0+

🔒 Security First

🎯 Best Practices

🔐 Never store raw PII
💾 Process in-memory only
🧹 Clear cache regularly
📝 Enable access logging
🔄 Regular updates

⚠️ Known Limitations

📏 Max 2048 tokens
🗣️ English-primary
💡 Domain adaptation needed
💾 GPU memory bound

📜 License

Apache License 2.0 • Made with ❤️ for Privacy

🤝 Support & Community

💬 LinkedIn
📧 Email Support