Instructions to use OPTML-Group/GradDiff-WMDP with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use OPTML-Group/GradDiff-WMDP with Transformers:

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

pipe = pipeline("text-generation", model="OPTML-Group/GradDiff-WMDP")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("OPTML-Group/GradDiff-WMDP")
model = AutoModelForCausalLM.from_pretrained("OPTML-Group/GradDiff-WMDP")
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 OPTML-Group/GradDiff-WMDP with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/OPTML-Group/GradDiff-WMDP

SGLang

How to use OPTML-Group/GradDiff-WMDP 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 "OPTML-Group/GradDiff-WMDP" \
    --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": "OPTML-Group/GradDiff-WMDP",
		"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 "OPTML-Group/GradDiff-WMDP" \
        --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": "OPTML-Group/GradDiff-WMDP",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use OPTML-Group/GradDiff-WMDP with Docker Model Runner:
```
docker model run hf.co/OPTML-Group/GradDiff-WMDP
```

GradDiff-WMDP / README.md

a-F1

Create README.md

0695dcc verified about 1 year ago

preview code

Raw

History Blame Contribute Delete

1.69 kB

	---
	license: mit
	datasets:
	- cais/wmdp
	language:
	- en
	base_model:
	- HuggingFaceH4/zephyr-7b-beta
	pipeline_tag: text-generation
	library_name: transformers
	tags:
	- unlearn
	- machine-unlearning
	- llm-unlearning
	- data-privacy
	- large-language-models
	- trustworthy-ai
	- trustworthy-machine-learning
	- language-model
	---

	# GradDiff-Unlearned Model on Task "WMDP"

	## Model Details

	- Unlearning:
	- Task: [🤗datasets/cais/wmdp wmdp-bio](https://huggingface.co/datasets/cais/wmdp)
	- Method: GradDiff
	- Smoothness Optimization: None
	- Origin Model: [🤗HuggingFaceH4/zephyr-7b-beta](https://huggingface.co/HuggingFaceH4/zephyr-7b-beta)
	- Code Base: [github.com/OPTML-Group/Unlearn-Smooth](https://github.com/OPTML-Group/Unlearn-Smooth)
	- Research Paper: ["Towards LLM Unlearning Resilient to Relearning Attacks: A Sharpness-Aware Minimization Perspective and Beyond"](https://arxiv.org/abs/2502.05374)

	## Loading the Model

	```python
	import torch
	from transformers import AutoModelForCausalLM

	model = AutoModelForCausalLM.from_pretrained("OPTML-Group/GradDiff-WMDP", torch_dtype=torch.bfloat16, trust_remote_code=True)
	```

	## Citation

	If you use this model in your research, please cite:
	```
	@article{fan2025towards,
	title={Towards LLM Unlearning Resilient to Relearning Attacks: A Sharpness-Aware Minimization Perspective and Beyond},
	author={Fan, Chongyu and Jia, Jinghan and Zhang, Yihua and Ramakrishna, Anil and Hong, Mingyi and Liu, Sijia},
	journal={arXiv preprint arXiv:2502.05374},
	year={2025}
	}
	```

	## Reporting Issues

	Reporting issues with the model: [github.com/OPTML-Group/Unlearn-Smooth](https://github.com/OPTML-Group/Unlearn-Smooth)