Reallm-Labs/Infi-MMR-3B

Inference

Our models are established on top of the Qwen2.5-VL family. So we include a simple use case here, and refer the readers to the standard inference procedure of Qwen2.5-VL.

from transformers import Qwen2_5_VLForConditionalGeneration, AutoProcessor
from qwen_vl_utils import process_vision_info

# default: Load the model on the available device(s)
model = Qwen2_5_VLForConditionalGeneration.from_pretrained(
    "Reallm-Labs/Infi-MMR-3B", torch_dtype="auto", device_map="auto"
)
min_pixels = 256*28*28
max_pixels = 1280*28*28
processor = AutoProcessor.from_pretrained("Reallm-Labs/Infi-MMR-3B", min_pixels=min_pixels, max_pixels=max_pixels)

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
text = processor.apply_chat_template(
    messages, tokenize=False, add_generation_prompt=True
)
image_inputs, video_inputs = process_vision_info(messages)
inputs = processor(
    text=[text],
    images=image_inputs,
    videos=video_inputs,
    padding=True,
    return_tensors="pt",
)
inputs = inputs.to(model.device)

# Inference: Generation of the output
generated_ids = model.generate(**inputs, max_new_tokens=4096)
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)

Citation Information

If you find this work useful, we would be grateful if you consider citing the following papers:

@article{liu2025infimmr,
  title={Infi-MMR: Curriculum-based Unlocking Multimodal Reasoning via Phased Reinforcement Learning in Multimodal Small Language Models},
  author={Zeyu Liu and Yuhang Liu and Guanghao Zhu and Congkai Xie and Zhen Li and Jianbo Yuan and Xinyao Wang and Qing Li and Shing-Chi Cheung and Shengyu Zhang and Fei Wu and Hongxia Yang},
  journal={arXiv preprint arXiv:2505.23091},
  year={2025}
}