Intel/gemma-4-26B-A4B-it-int4-AutoRound

Model Details

This model is an int4 model with group_size 128 and symmetric quantization of google/gemma-4-26B-A4B-it generated by intel/auto-round. Please follow the license of the original model.

Transformes inference

from transformers import AutoProcessor, AutoModelForCausalLM

MODEL_ID = "Intel/gemma-4-26B-A4B-it-int4-AutoRound/"
# Load model
processor = AutoProcessor.from_pretrained(MODEL_ID)
model = AutoModelForCausalLM.from_pretrained(
    MODEL_ID,
    dtype="auto",
    device_map="auto"
)


# Prompt - add image before text
messages = [
    {
        "role": "user", "content": [
            {"type": "image", "url": "https://raw.githubusercontent.com/google-gemma/cookbook/refs/heads/main/Demos/sample-data/GoldenGate.png"},
            {"type": "text", "text": "What is shown in this image?"}
        ]
    }
]

# Process input
inputs = processor.apply_chat_template(
    messages,
    tokenize=True,
    return_dict=True,
    return_tensors="pt",
    add_generation_prompt=True,
).to(model.device)
input_len = inputs["input_ids"].shape[-1]

# Generate output
outputs = model.generate(**inputs, max_new_tokens=512)
response = processor.decode(outputs[0][input_len:], skip_special_tokens=False)

# Parse output
print(processor.parse_response(response))

{'content': 'An outdoor, daytime, medium shot shows the Golden Gate Bridge spanning a body of water, with a large, tan building in the foreground and hills in the distance. The bridge is a deep red color and is supported by two large towers. The water is a calm, dark blue, and there are some small waves. The building in the foreground is made of tan bricks and has several windows. In the foreground, there is a large rock in the water, and a small bird is perched on top of it. The sky is a clear, light blue. The sun is shining brightly, casting a soft light on the scene.', 'role': 'assistant'}

Generate the Model

https://github.com/intel/auto-round/pull/1656 is required

auto-round
/models/gemma-4-26B-A4B-it
--output_dir
"./quantized"
--scheme
W4A16

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

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