DAS1.1 下面的 paddlepaddle-2.5.2_dtk2404_git7ad272b3-cp310-cp310-manylinux2014_x86_64.whl 使用 PaddleYOLO，出现 muticlass_nms3 的相关问题。

复现方法：

```python

git clone https://github.com/PaddlePaddle/PaddleYOLO

cd PaddleYOLO

pip install -r requirements.txt

python tools/infer.py -c configs/yolov8/yolov8_n_500e_coco.yml -o use_gpu=true weights=https://paddledet.bj.bcebos.com/models/yolov8_n_500e_coco.pdparams --infer_img=demo/000000014439.jpg

```

解决方案如下:

更改如下 PaddleYOLO/ppdet/modeling/ops.py 文件:

1. 强制使用非legacy模式：

```python

# try:

# import paddle._legacy_C_ops as C_ops

# except:

# import paddle._C_ops as C_ops

import paddle._C_ops as C_ops

```

2. 更改`def multiclass_nms` 函数内部`attrs`内容：

```python

# attrs = ('background_label', background_label, 'score_threshold',

# score_threshold, 'nms_top_k', nms_top_k, 'nms_threshold',

# nms_threshold, 'keep_top_k', keep_top_k, 'nms_eta', nms_eta,

# 'normalized', normalized)

attrs = (

score_threshold,

nms_top_k,

keep_top_k,

nms_threshold,

normalized,

nms_eta,

background_label,

)