Instructions to use intel-ai/Qwen3.8-Flash-Next-MXFP4-BF16SharedExperts-CT-AutoRound with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use intel-ai/Qwen3.8-Flash-Next-MXFP4-BF16SharedExperts-CT-AutoRound with Transformers:

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

pipe = pipeline("text-generation", model="intel-ai/Qwen3.8-Flash-Next-MXFP4-BF16SharedExperts-CT-AutoRound")
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)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoProcessor, AutoModelForMultimodalLM

processor = AutoProcessor.from_pretrained("intel-ai/Qwen3.8-Flash-Next-MXFP4-BF16SharedExperts-CT-AutoRound")
model = AutoModelForMultimodalLM.from_pretrained("intel-ai/Qwen3.8-Flash-Next-MXFP4-BF16SharedExperts-CT-AutoRound", 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=256)
print(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use intel-ai/Qwen3.8-Flash-Next-MXFP4-BF16SharedExperts-CT-AutoRound with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "intel-ai/Qwen3.8-Flash-Next-MXFP4-BF16SharedExperts-CT-AutoRound"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "intel-ai/Qwen3.8-Flash-Next-MXFP4-BF16SharedExperts-CT-AutoRound",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/intel-ai/Qwen3.8-Flash-Next-MXFP4-BF16SharedExperts-CT-AutoRound

SGLang

How to use intel-ai/Qwen3.8-Flash-Next-MXFP4-BF16SharedExperts-CT-AutoRound 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 "intel-ai/Qwen3.8-Flash-Next-MXFP4-BF16SharedExperts-CT-AutoRound" \
    --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": "intel-ai/Qwen3.8-Flash-Next-MXFP4-BF16SharedExperts-CT-AutoRound",
		"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 "intel-ai/Qwen3.8-Flash-Next-MXFP4-BF16SharedExperts-CT-AutoRound" \
        --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": "intel-ai/Qwen3.8-Flash-Next-MXFP4-BF16SharedExperts-CT-AutoRound",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use intel-ai/Qwen3.8-Flash-Next-MXFP4-BF16SharedExperts-CT-AutoRound with Docker Model Runner:
```
docker model run hf.co/intel-ai/Qwen3.8-Flash-Next-MXFP4-BF16SharedExperts-CT-AutoRound
```

Model Details

This model is an MXFP4 (mixed MXFP4 experts / MXFP8 linear) model of Qwen/Qwen3.8-Flash-Next generated by intel/auto-round using RTN. Please follow the license of the original model.

Quantized (MXFP4): routed experts
Kept in BF16: attention, router, embeddings, lm_head, vision tower, shared expert, MTP, PLE table

How to Use

Environment

uv pip install --index-strategy unsafe-best-match \
    --extra-index-url https://wheels.vllm.ai/9815732c16ca076c60b65fccd787167d6a9c5328/cu130 \
    "vllm==0.29.1rc1.dev528+g9815732c1"
python - <<'PY'
from importlib.util import find_spec
from pathlib import Path
import subprocess
from huggingface_hub import hf_hub_download

spec = find_spec("vllm")
if spec is None or spec.origin is None:
    raise RuntimeError("Install the pinned vLLM build first")
root = Path(spec.origin).parent.parent
for filename in ['patches/qwen38fn-ct-ple-dev528.patch']:
    patch = hf_hub_download("intel-ai/Qwen3.8-Flash-Next-MXFP4-BF16SharedExperts-CT-AutoRound", filename, revision="415d9ae4aa48e6067a3f0d2ed5e10a932e4f123e")
    command = ["patch", "--batch", "--fuzz=0", "-p1", "-d", str(root), "-i", patch]
    reverse = subprocess.run(command + ["--reverse", "--dry-run"], capture_output=True)
    if reverse.returncode == 0:
        print(filename + ": already applied")
        continue
    subprocess.run(command + ["--forward", "--dry-run"], check=True)
    subprocess.run(command + ["--forward"], check=True)
PY

vLLM Usage

vllm serve intel-ai/Qwen3.8-Flash-Next-MXFP4-BF16SharedExperts-CT-AutoRound \
    --reasoning-parser qwen3

Evaluation

Benchmark	BF16	MXFP4	Recovery
GSM8K	96.74	96.66	99.92%
MMLU	86.52	86.43	99.90%
PIQA	81.94	81.45	99.40%
HellaSwag	69.27	68.32	98.63%
Average	83.62	83.22	99.52%

Generate the Model

auto-round \
    --model_name Qwen/Qwen3.8-Flash-Next \
    --model_free \
    --scheme MXFP8 \
    --ignore_layers "visual,lm_head,embed_tokens,mlp.gate,mlp.shared_expert_gate,in_proj_a,in_proj_b,block_inject_weight,ple,mtp,hyper_connection,indexer,shared_expert" \
    --layer_config "{mlp.experts:{bits:4,data_type:mx_fp}}" \
    --format llm_compressor \
    --output_dir Qwen3.8-Flash-Next-MXFP4-BF16SharedExperts-CT-AutoRound

Ethical Considerations and Limitations

The model can produce factually incorrect output, and should not be relied on to produce factually accurate information. Because of the limitations of the pretrained model and the finetuning datasets, it is possible that this model could generate lewd, biased or otherwise offensive outputs. Therefore, before deploying any applications of the model, developers should perform safety testing.

Caveats and Recommendations

Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. Here are a couple of useful links to learn more about Intel's AI software:

Disclaimer

The license on this model does not constitute legal advice. We are not responsible for the actions of third parties who use this model. Please consult an attorney before using this model for commercial purposes.

Cite

@article{cheng2023optimize,
  title={Optimize weight rounding via signed gradient descent for the quantization of llms},
  author={Cheng, Wenhua and Zhang, Weiwei and Shen, Haihao and Cai, Yiyang and He, Xin and Lv, Kaokao and Liu, Yi},
  journal={arXiv preprint arXiv:2309.05516},
  year={2023}
}

arxiv github

Signed-off-by: Lv, Kaokao kaokao.lv@intel.com

Downloads last month: 340

Model tree for intel-ai/Qwen3.8-Flash-Next-MXFP4-BF16SharedExperts-CT-AutoRound

Base model

Qwen/Qwen3.8-Flash-Next

Quantized

(530)

this model

Paper for intel-ai/Qwen3.8-Flash-Next-MXFP4-BF16SharedExperts-CT-AutoRound

Optimize Weight Rounding via Signed Gradient Descent for the Quantization of LLMs

Paper • 2309.05516 • Published Sep 11, 2023 • 16