推理响应的数据是一串json数据

[hackerhaiJu]

import os  # 添加os模块用于路径检查
import torch
from modelscope import snapshot_download
from transformers import Qwen2VLForConditionalGeneration, AutoTokenizer, AutoProcessor
from qwen_vl_utils import process_vision_info

os.environ["CUDA_VISIBLE_DEVICES"] = "2,3"

def run_model(model_path: str):
    # 检查模型路径是否存在
    if not os.path.exists(model_path):
        raise FileNotFoundError(f"Model path {model_path} does not exist.")

    model = Qwen2VLForConditionalGeneration.from_pretrained(
        model_path, device_map="auto", torch_dtype='auto'
    )
    processor = AutoProcessor.from_pretrained(model_path)

    messages = [
        {
            "role": "user",
            "content": [
                {
                    "type": "image",
                    "image": "file:///app/qwen/images/demo3.png",
                },
                {"type": "text", "text": "识别图中的表单信息，并且以key:value的形式返回给我"},
            ],
        }
    ]
    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("cuda")

    # 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)


if __name__ == '__main__':
    run_model('/app/qwen/Qwen2-VL-7B-Instruct')

为什么推理出的是一串json数据呢？