RT-DETR-L_wireless_table_cell_det_infer 模型修改推理图像尺度出现报错

[zhujun5164]

Checklist:

• 查找历史相关issue寻求解答
• 翻阅FAQ
• 翻阅PaddleX 文档
• 确认bug是否在新版本里还未修复

描述问题

我使用paddlex里的RT-DETR-L_wireless_table_cell_det_infer模型，在create_model的时候尝试修改img_size参数，运行推理后出现报错

复现

`
from paddlex import create_model

model = create_model(model_name="RT-DETR-L_wireless_table_cell_det", img_size=1280)
output = model.predict("d:/test.png", batch_size=1)
for res in output:
res.print(json_format=False)
res.save_to_img("./output/")
res.save_to_json("./output/res.json")
`
出现报错如下
self.predictor.run()
ValueError: In user code:

InvalidArgumentError: Broadcast dimension mismatch. Operands could not be broadcast together with the shape of X = [1, 1600, 256] and the shape of Y = [1, 400, 256]. Received [1600] in X is not equal to [400] in Y at i:1.
  [Hint: Expected x_dims_array[i] == y_dims_array[i] || (x_dims_array[i] <= 1 && x_dims_array[i] != 0) || (y_dims_array[i] <= 1 && y_dims_array[i] != 0) == true, but received x_dims_array[i] == y_dims_array[i] || (x_dims_array[i] <= 1 && x_dims_array[i] != 0) || (y_dims_array[i] <= 1 && y_dims_array[i] != 0):0 != true:1.] (at ..\paddle/phi/kernels/funcs/common_shape.h:86)

环境

1. 请提供您使用的PaddlePaddle、PaddleX版本号、Python版本号
   python 3.11.1
   paddlepaddle 3.2.1 gpu
   paddlex 3.3.8

2. 请提供您使用的操作系统信息，如Linux/Windows/MacOS
   window

3. 请问您使用的CUDA/cuDNN的版本号是？