Instructions to use SwissNeuron/Qwen3.8-27B-SwissNeuron-Derisked-NVFP4 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use SwissNeuron/Qwen3.8-27B-SwissNeuron-Derisked-NVFP4 with Transformers:

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

pipe = pipeline("image-text-to-text", model="SwissNeuron/Qwen3.8-27B-SwissNeuron-Derisked-NVFP4")
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("SwissNeuron/Qwen3.8-27B-SwissNeuron-Derisked-NVFP4")
model = AutoModelForMultimodalLM.from_pretrained("SwissNeuron/Qwen3.8-27B-SwissNeuron-Derisked-NVFP4", 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 SwissNeuron/Qwen3.8-27B-SwissNeuron-Derisked-NVFP4 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "SwissNeuron/Qwen3.8-27B-SwissNeuron-Derisked-NVFP4"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "SwissNeuron/Qwen3.8-27B-SwissNeuron-Derisked-NVFP4",
		"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/SwissNeuron/Qwen3.8-27B-SwissNeuron-Derisked-NVFP4

SGLang

How to use SwissNeuron/Qwen3.8-27B-SwissNeuron-Derisked-NVFP4 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 "SwissNeuron/Qwen3.8-27B-SwissNeuron-Derisked-NVFP4" \
    --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": "SwissNeuron/Qwen3.8-27B-SwissNeuron-Derisked-NVFP4",
		"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 "SwissNeuron/Qwen3.8-27B-SwissNeuron-Derisked-NVFP4" \
        --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": "SwissNeuron/Qwen3.8-27B-SwissNeuron-Derisked-NVFP4",
		"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 SwissNeuron/Qwen3.8-27B-SwissNeuron-Derisked-NVFP4 with Docker Model Runner:
```
docker model run hf.co/SwissNeuron/Qwen3.8-27B-SwissNeuron-Derisked-NVFP4
```

Qwen3.8-27B-SwissNeuron-Derisked-NVFP4

NVIDIA NVFP4 quantization of SwissNeuron/Qwen3.8-27B-SwissNeuron-Derisked.

Source revision: repaired direct-answer SFT + fresh latest-SFT DWM α=0.1
Format: NVIDIA ModelOpt unified Hugging Face checkpoint
Language backbone: NVFP4 W4A4 with calibrated block scales
Vision tower and MTP draft head remain BF16
Intended hardware: NVIDIA Blackwell (RTX 50 series, RTX PRO 6000 Blackwell, B200/B300, GB200/GB300)
Active config retains factor-4 YaRN to 1,048,576 tokens

Runtime

Use current vLLM or TensorRT-LLM with ModelOpt NVFP4 support. Blackwell is required for native FP4 Tensor Core execution.

vllm serve SwissNeuron/Qwen3.8-27B-SwissNeuron-Derisked-NVFP4 \
  --trust-remote-code

Thinking modes

Both native thinking and non-thinking modes are available through the included Qwen chat template. Set enable_thinking=False only when a direct lower-latency response is desired.

Limitations

NVFP4 is hardware-specific. Do not expect native acceleration on pre-Blackwell GPUs. Re-evaluate critical workloads and long-context retrieval rather than assuming BF16 parity.

Downloads last month: 17

Safetensors

Model size

15B params

Tensor type

BF16

Model tree for SwissNeuron/Qwen3.8-27B-SwissNeuron-Derisked-NVFP4

Base model

Qwen/Qwen3.8-27B

Finetuned

SwissNeuron/Qwen3.8-27B-SwissNeuron-Derisked

Quantized

(4)

this model