Instructions to use apolo13x/Qwen3.5-27B-quantized.w4a16 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use apolo13x/Qwen3.5-27B-quantized.w4a16 with Transformers:

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

pipe = pipeline("image-text-to-text", model="apolo13x/Qwen3.5-27B-quantized.w4a16")
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("apolo13x/Qwen3.5-27B-quantized.w4a16")
model = AutoModelForMultimodalLM.from_pretrained("apolo13x/Qwen3.5-27B-quantized.w4a16", 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 apolo13x/Qwen3.5-27B-quantized.w4a16 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "apolo13x/Qwen3.5-27B-quantized.w4a16"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "apolo13x/Qwen3.5-27B-quantized.w4a16",
		"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/apolo13x/Qwen3.5-27B-quantized.w4a16

SGLang

How to use apolo13x/Qwen3.5-27B-quantized.w4a16 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 "apolo13x/Qwen3.5-27B-quantized.w4a16" \
    --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": "apolo13x/Qwen3.5-27B-quantized.w4a16",
		"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 "apolo13x/Qwen3.5-27B-quantized.w4a16" \
        --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": "apolo13x/Qwen3.5-27B-quantized.w4a16",
		"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 apolo13x/Qwen3.5-27B-quantized.w4a16 with Docker Model Runner:
```
docker model run hf.co/apolo13x/Qwen3.5-27B-quantized.w4a16
```

Qwen3.5-27B-quantized.w4a16

This is a quantized version of Qwen/Qwen3.5-27B. This model accepts text and images as inputs and generates text as outputs. The weights were quantized to INT4 using GPTQ via llm-compressor with 512 calibration samples from HuggingFaceH4/ultrachat_200k, reducing the model size from 51.8 GB to 17.3 GB (~3.0x reduction) while maintaining 100.3% average accuracy recovery.

Inference

As of 2/27/2026, this model is supported in vLLM nightly. To serve the model:

vllm serve Kbenkhaled/Qwen3.5-27B-quantized.w4a16 \
    --reasoning-parser qwen3 \
    --enable-prefix-caching

Evaluation

Evaluated with lm-evaluation-harness, 0-shot, thinking mode ON.

Benchmark	Qwen3.5-27B	Qwen3.5-27B-quantized.w4a16 (this model)	Recovery
GPQA Diamond	80.30%	80.81%	100.6%
IFEval	95.08%	95.20%	100.1%
MMLU-Redux	93.90%	94.13%	100.2%
Average	89.76%	90.05%	100.3%

Downloads last month: 1,648

Safetensors

Model size

28B params

Tensor type

I64

I32

BF16

Model tree for apolo13x/Qwen3.5-27B-quantized.w4a16

Base model

Qwen/Qwen3.5-27B

Quantized

(230)

this model

apolo13x
/

Qwen3.5-27B-quantized.w4a16

Qwen3.5-27B-quantized.w4a16

Inference

Evaluation

Model tree for apolo13x/Qwen3.5-27B-quantized.w4a16

Dataset used to train apolo13x/Qwen3.5-27B-quantized.w4a16