Instructions to use ArliAI/Qwen3.5-27B-Derestricted with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use ArliAI/Qwen3.5-27B-Derestricted with Transformers:

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

pipe = pipeline("text-generation", model="ArliAI/Qwen3.5-27B-Derestricted")
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, AutoModelForImageTextToText

processor = AutoProcessor.from_pretrained("ArliAI/Qwen3.5-27B-Derestricted")
model = AutoModelForImageTextToText.from_pretrained("ArliAI/Qwen3.5-27B-Derestricted")
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 ArliAI/Qwen3.5-27B-Derestricted with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "ArliAI/Qwen3.5-27B-Derestricted"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ArliAI/Qwen3.5-27B-Derestricted",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/ArliAI/Qwen3.5-27B-Derestricted

SGLang

How to use ArliAI/Qwen3.5-27B-Derestricted 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 "ArliAI/Qwen3.5-27B-Derestricted" \
    --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": "ArliAI/Qwen3.5-27B-Derestricted",
		"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 "ArliAI/Qwen3.5-27B-Derestricted" \
        --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": "ArliAI/Qwen3.5-27B-Derestricted",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use ArliAI/Qwen3.5-27B-Derestricted with Docker Model Runner:
```
docker model run hf.co/ArliAI/Qwen3.5-27B-Derestricted
```

Still quite restricted for a "derestricted" claim.

by BigBeavis - opened Mar 10

Discussion

BigBeavis

Mar 10

Sure, the model doesn't refuse a request outright (sometimes), but instead it just quickly switches topics (quite jarringly at that) before it has to deal with any morally grey stuff.

AutisticPancake

Mar 19

Give it a system prompt that encourages it to be more direct, unflinchingly explicit and free of restraints - it will likely listen and change its behavior (while the original model wouldn't).

Models treated with Norm-Preserving Biprojected Abliteration tend to retain "soft refusals", necessary for mimicking certain behaviours (such as confrontational, irritable, meek, etc.).

The kind of "derestriction" you originally expected - when LLM immediately agrees with anything and everything - is an indication of significant damage done to its 'brain'.

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment