Instructions to use webhie/Qwen3.6-27B-int4-AutoRound-Code with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use webhie/Qwen3.6-27B-int4-AutoRound-Code with Transformers:

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

pipe = pipeline("image-text-to-text", model="webhie/Qwen3.6-27B-int4-AutoRound-Code")
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("webhie/Qwen3.6-27B-int4-AutoRound-Code")
model = AutoModelForImageTextToText.from_pretrained("webhie/Qwen3.6-27B-int4-AutoRound-Code")
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 webhie/Qwen3.6-27B-int4-AutoRound-Code with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "webhie/Qwen3.6-27B-int4-AutoRound-Code"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "webhie/Qwen3.6-27B-int4-AutoRound-Code",
		"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/webhie/Qwen3.6-27B-int4-AutoRound-Code

SGLang

How to use webhie/Qwen3.6-27B-int4-AutoRound-Code 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 "webhie/Qwen3.6-27B-int4-AutoRound-Code" \
    --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": "webhie/Qwen3.6-27B-int4-AutoRound-Code",
		"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 "webhie/Qwen3.6-27B-int4-AutoRound-Code" \
        --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": "webhie/Qwen3.6-27B-int4-AutoRound-Code",
		"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 webhie/Qwen3.6-27B-int4-AutoRound-Code with Docker Model Runner:
```
docker model run hf.co/webhie/Qwen3.6-27B-int4-AutoRound-Code
```

Qwen3.6-27B-int4-AutoRound-Code / README.md

edwinbrowwn

Upload folder using huggingface_hub (#1)

2264cf0 24 days ago

preview code

raw

history blame contribute delete

9.65 kB

	---
	license: apache-2.0
	license_link: https://huggingface.co/Qwen/Qwen3.6-27B/blob/main/LICENSE
	base_model: Qwen/Qwen3.6-27B
	base_model_relation: quantized
	pipeline_tag: image-text-to-text
	library_name: transformers
	tags:
	- qwen3_5
	- autoround
	- int4
	- w4g128
	- w4a16
	- quantization
	- vllm
	- multimodal
	- mtp
	- speculative-decoding
	- code
	- coding
	---
	# Qwen3.6-27B INT4 AutoRound — Code Calibrated (Best Recipe)

	A W4A16 (INT4 weight, FP16 activation) quantization of [`Qwen/Qwen3.6-27B`](https://huggingface.co/Qwen/Qwen3.6-27B), produced with [Intel's AutoRound](https://github.com/intel/auto-round).

	> Key difference from the standard AutoRound quant: This variant was calibrated on a normalized and sampled subset of [nvidia/OpenCodeInstruct](https://huggingface.co/datasets/nvidia/OpenCodeInstruct) — a ~5 M sample, execution-verified coding dataset — instead of the default general-purpose pile corpus. Calibrating on domain-specific data guides AutoRound's weight-rounding optimization to minimize quantization error on the token distributions that matter most for code, improving accuracy on code generation, reasoning, and instruction-following for programming tasks. The `auto-round-best` preset was used (1000 iterations, 512 calibration samples), which runs ~4–5× slower than the standard recipe but achieves the best possible INT4 accuracy. MTP (speculative decoding) and image/vision inputs work out of the box with no post-processing required.

	## TL;DR

	- Base: Qwen3.6-27B (27B dense VLM)
	- Quant: INT4 W4A16, group_size 128, symmetric
	- Tool: `auto-round-best` (1000 iters, 512 samples, torch.compile)
	- Calibration dataset: `nvidia/OpenCodeInstruct` (coding domain)
	- Size: ~18 GB (down from ~54 GB BF16) — 3× reduction
	- MTP: Native Multi-Token Prediction head preserved in BF16 — enables native speculative decoding in vLLM (~85–90% draft acceptance, ~2× throughput)
	- Vision: Image inputs work via the MoonViT encoder (weights kept at original BF16/FP16 precision)

	## Why code calibration?

	AutoRound's algorithm optimizes weight rounding by minimizing the difference between the quantized model's outputs and the full-precision model's outputs on a set of calibration samples. The calibration dataset therefore shapes which activations and weight patterns are prioritized during optimization.

	Using a normalized and sampled subset of `nvidia/OpenCodeInstruct` — a large, execution-verified dataset of coding problems and solutions — means the rounding decisions are tuned for code-style token distributions: identifiers, keywords, indentation patterns, and structured reasoning. In practice this tends to:

	- Better preserve accuracy on code generation benchmarks relative to a pile-calibrated quant
	- Improve instruction following for programming tasks (function signatures, docstrings, tool calls)
	- Retain structured output quality (JSON, markdown code blocks, structured diffs)

	If your primary use-case is code generation or an AI coding assistant, this variant is the recommended choice. For general-purpose or multimodal usage, see the standard [`Qwen3.6-27B-int4-AutoRound`](https://huggingface.co/webhie/Qwen3.6-27B-int4-AutoRound) quant.

	## Quick inference with vLLM (with MTP speculative decoding)

	Requires vLLM v0.19.1+ with Qwen3_5 MTP support. Set the following environment variables before starting:

	```bash
	export VLLM_USE_FLASHINFER_SAMPLER=1
	export VLLM_ALLOW_LONG_MAX_MODEL_LEN=1
	export VLLM_FLOAT32_MATMUL_PRECISION=high
	export PYTORCH_CUDA_ALLOC_CONF="expandable_segments:True,max_split_size_mb:512"
	export VLLM_NO_USAGE_STATS=1
	export VLLM_MEMORY_PROFILER_ESTIMATE_CUDAGRAPHS=1
	export VLLM_MARLIN_USE_ATOMIC_ADD=1
	export OMP_NUM_THREADS=1
	export CUDA_DEVICE_MAX_CONNECTIONS=8
	export NCCL_CUMEM_ENABLE=0
	export NCCL_P2P_DISABLE=1
	```

	```bash
	vllm serve webhie/Qwen3.6-27B-int4-AutoRound-Code \
	--served-model-name qwen3.6-27b \
	--host 0.0.0.0 --port 11434 \
	--trust-remote-code \
	--dtype auto \
	--quantization auto_round \
	--max-model-len 200704 \
	--gpu-memory-utilization 0.92 \
	--max-num-seqs 4 \
	--kv-cache-dtype fp8_e4m3 \
	--attention-backend flashinfer \
	--performance-mode throughput \
	--max-num-batched-tokens 2048 \
	--enable-chunked-prefill \
	--enable-auto-tool-choice \
	--tool-call-parser qwen3_coder \
	--reasoning-parser qwen3 \
	--default-chat-template-kwargs '{"preserve_thinking":true}' \
	--override-generation-config '{"temperature":0.6,"top_p":0.95,"top_k":20,"min_p":0.0,"presence_penalty":0.0,"repetition_penalty":1.0}' \
	--enable-prompt-tokens-details \
	--speculative-config '{"method":"mtp","num_speculative_tokens":3}'
	```

	Remove `--speculative-config` to disable MTP speculative decoding. See the [vllm-blackwell-guide](https://github.com/lastloop-ai/vllm-blackwell-guide) repo for a full Docker Compose setup with all env vars pre-configured.

	### OpenAI-compatible request

	```python
	from openai import OpenAI
	client = OpenAI(base_url="http://localhost:11434/v1", api_key="EMPTY")
	r = client.chat.completions.create(
	model="qwen3.6-27b",
	messages=[{"role": "user", "content": "Write a quicksort in Python."}],
	max_tokens=512,
	)
	print(r.choices[0].message.content)
	```

	### Transformers (no spec decoding)

	```python
	from transformers import AutoModelForCausalLM, AutoTokenizer
	m = AutoModelForCausalLM.from_pretrained(
	"webhie/Qwen3.6-27B-int4-AutoRound-Code",
	trust_remote_code=True,
	device_map="auto",
	)
	tok = AutoTokenizer.from_pretrained("webhie/Qwen3.6-27B-int4-AutoRound-Code")
	msg = [{"role": "user", "content": "Write a binary search in Python."}]
	ids = tok.apply_chat_template(msg, add_generation_prompt=True, return_tensors="pt").to(m.device)
	print(tok.decode(m.generate(ids, max_new_tokens=256)[0]))
	```

	## Quantization details

	\| Field \| Value \|
	\|---\|---\|
	\| Base \| `Qwen/Qwen3.6-27B` \|
	\| Method \| AutoRound (`intel/auto-round`), best recipe \|
	\| Scheme \| W4A16 (4-bit weights, FP16 activations) \|
	\| Bits \| 4 \|
	\| Group size \| 128 \|
	\| Symmetric \| yes \|
	\| Packing format \| `auto_round:auto_gptq` \|
	\| Unquantized layers \| `linear_attn.in_proj_a/b`, all LayerNorms, RMSNorms, router gates \|
	\| Calibration dataset \| Normalized & sampled subset of [`nvidia/OpenCodeInstruct`](https://huggingface.co/datasets/nvidia/OpenCodeInstruct) \|
	\| Calibration samples \| 512 \|
	\| Iterations \| 1000 \|
	\| torch.compile \| enabled \|
	\| GPU used for quant \| 1× RTX 5090 (32 GB, SM120), `low_gpu_mem_usage=True` \|

	### Unquantized layers — why

	- `linear_attn.in_proj_a/b`: low-rank projections in Qwen3.6's Gated DeltaNet whose shapes aren't divisible by 32 (group_size), so AutoRound skips them automatically. Tiny fraction of total parameters.
	- Norms, routers: precision-sensitive and very small — kept at full precision.

	## Performance

	Benchmarked on 1× RTX 5090 (32 GB) with vLLM + FP8 KV cache + MTP n=3:

	\| Config \| Throughput \|
	\|---\|---:\|
	\| vLLM + MTP n=3 \| ~150 tok/s \|
	\| vLLM (MTP disabled) \| ~70 tok/s \|

	The ~2× speedup comes from ~85–90% draft acceptance via MTP speculative decoding with `num_speculative_tokens: 3`.

	## Reproduction

	```bash
	pip install auto-round

	# The calibration data was first normalized and sampled from nvidia/OpenCodeInstruct
	# (formatting cleaned, deduplicated, balanced across domains) and exported as a
	# local JSON file before quantization. Pass your own prepared subset with:
	# --dataset ./subset_10k.json

	auto-round-best \
	--model Qwen/Qwen3.6-27B \
	--scheme W4A16 \
	--format auto_round \
	--output_dir Qwen3.6-27B-int4-AutoRound-Code \
	--enable_torch_compile \
	--low_gpu_mem_usage \
	--device_map 0
	```

	No post-processing needed — MTP and image inputs work out of the box.

	## Acknowledgements

	- [Alibaba / Qwen team](https://huggingface.co/Qwen) for the base [Qwen3.6-27B](https://huggingface.co/Qwen/Qwen3.6-27B) model
	- [Intel AutoRound](https://github.com/intel/auto-round) team for the quantization framework and the `auto-round-best` recipe
	- [NVIDIA](https://huggingface.co/nvidia) for the [OpenCodeInstruct](https://huggingface.co/datasets/nvidia/OpenCodeInstruct) calibration dataset — ~5 M execution-verified coding samples used to domain-adapt this quant
	- [Lorbus](https://huggingface.co/Lorbus) for the original AutoRound quant of this model that inspired this release
	- [@eugr](https://github.com/eugr) for the [spark-vllm-docker](https://github.com/eugr/spark-vllm-docker) fork and TurboQuant KV cache work
	- [vLLM project](https://github.com/vllm-project/vllm) for the inference engine and Qwen3_5 MTP support

	## License

	Apache 2.0 — same as [Qwen3.6-27B base](https://huggingface.co/Qwen/Qwen3.6-27B).

	## Citation

	If you use this quant, please cite the original Qwen3.6 release (see base model card), the AutoRound paper, and the OpenCodeInstruct dataset:

	```bibtex
	@article{cheng2023autoround,
	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}
	}

	@misc{nvidia2025opencode,
	title = {OpenCodeInstruct: A Large-scale Instruction Tuning Dataset for Code LLMs},
	author = {NVIDIA},
	year = {2025},
	url = {https://huggingface.co/datasets/nvidia/OpenCodeInstruct}
	}
	```