You can deploy this model and start earning money today!

Intel/
Qwen3-VL-30B-A3B-Instruct-int4-AutoRound

Name: Qwen3-VL-30B-A3B-Instruct-int4-AutoRound
Availability: InStock

safetensors

qwen3_vl_moe

2309.05516

Qwen/Qwen3-VL-30B-A3B-Instruct

quantized

4-bit

auto-round

Views

No views yet

Model Details

This model is an int4 model with group_size 128 and symmetric quantization of Qwen/Qwen3-VL-30B-A3B-Instruct generated by intel/auto-round. Please refer to Section Generate the model for more details. Please follow the license of the original model.

How To Use

INT4 Inference

Transformers (CPU/Intel GPU/CUDA)

Please make sure you are using auto_round version ≥ 0.9.6

python

from transformers import Qwen3VLMoeForConditionalGeneration, AutoProcessor

# default: Load the model on the available device(s)
model_name = "Intel/Qwen3-VL-30B-A3B-Instruct-int4-AutoRound"
model = Qwen3VLMoeForConditionalGeneration.from_pretrained(
    model_name, dtype="auto", device_map="auto"
)

processor = AutoProcessor.from_pretrained(model_name)

messages = [
    {
        "role": "user",
        "content": [
            {
                "type": "image",
                "image": "https://qianwen-res.oss-cn-beijing.aliyuncs.com/Qwen-VL/assets/demo.jpeg",
            },
            {"type": "text", "text": "Describe this image."},
        ],
    }
]

# Preparation for inference
inputs = processor.apply_chat_template(
    messages,
    tokenize=True,
    add_generation_prompt=True,
    return_dict=True,
    return_tensors="pt"
)

# Inference: Generation of the output
generated_ids = model.generate(**inputs, max_new_tokens=128)
generated_ids_trimmed = [
    out_ids[len(in_ids) :] for in_ids, out_ids in zip(inputs.input_ids, generated_ids)
]
output_text = processor.batch_decode(
    generated_ids_trimmed, skip_special_tokens=True, clean_up_tokenization_spaces=False
)
print(output_text)
"""
This is a heartwarming and serene photograph capturing a moment of connection between a woman and her dog on a beach at sunset.
\n\nThe image is taken from a side angle, focusing on the interaction between the two subjects. A woman with long, dark hair is sitting on the sand,
turned towards a large, light-colored dog, likely a yellow Labrador Retriever. The dog is sitting upright and is actively engaging with the woman,
extending its right front paw to give her a "high five." The woman is smiling warmly, looking at the dog with affection as she reciprocates the gesture.
\n\nThe setting is a sandy beach with gentle
"""

vLLM (CPU/Intel GPU/CUDA)

start a vllm server:

bash

python3 -m vllm.entrypoints.openai.api_server \
    --host localhost \
    --port 4321 \
    --trust-remote-code \
    --model Intel/Qwen3-VL-30B-A3B-Instruct-int4-AutoRound \
    --tensor-parallel-size 1

send request:

bash

curl --noproxy '*' http://127.0.0.1:4321/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{
    "model": "Intel/Qwen3-VL-30B-A3B-Instruct-int4-AutoRound",
    "messages": [
      {
        "role": "user",
        "content": [
          { "type": "text", "text": "Describe this image." },
          {
            "type": "image_url",
            "image_url": {
              "url": "https://qianwen-res.oss-cn-beijing.aliyuncs.com/Qwen-VL/assets/demo.jpeg"
            }
          }
        ]
      }
    ],
    "max_tokens": 256,
    "temperature": 0.6
  }'

"""
This is a heartwarming, full-shot photograph capturing a joyful moment between a woman and her dog on a beach at sunset.
\n\n- **Subjects and Interaction**: The central focus is a woman and a light-colored Labrador Retriever sitting on the sand.
 The woman, with long dark hair, is smiling warmly as she looks at her dog. The dog, wearing a blue harness with a colorful pattern,
 is sitting upright and has its right paw raised to meet the woman's hand in a \"high-five\" gesture. The woman is holding a small treat in her other hand,
 suggesting a moment of training or play.\n- **Setting and Atmosphere**: The scene is set on a wide, sandy beach with the ocean in the background.
 A gentle wave is breaking on the shore. The lighting is characteristic of a golden hour, with the sun low on the horizon,
 casting a warm, soft glow across the scene and creating a lens flare on the right side of the image. The sky is bright and hazy,
 indicating the time is either sunrise or sunset.\n- **Composition and Mood**: The woman and dog are positioned in the center of the frame,
 creating a balanced and intimate composition. The overall mood is one of happiness, companionship, and peaceful connection with nature.
"""

Generate the model

Please make sure you are using auto_round version ≥ 0.9.5

bash

auto_round \
--model=Qwen/Qwen3-VL-30B-A3B-Instruct \
--fp_layers=mlp.gate \
--format=auto_round \
--output_dir=./tmp_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:

Intel Neural Compressor link

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