自己调用yolov5模型进行前向推理时的报错

当我在自己的工程中调用yolov5的目标检测模型进行推理，代码大致如图：


当运行到如图箭头所指的位置的时候报如下错误：

Traceback (most recent call last):
File “/home/yons/train/code/mmpose/inference.py”, line 81, in
pred = yolo_model(img)[0] # 进行预测
File “/home/yons/miniconda3/envs/openmmlab/lib/python3.8/site-packages/torch/nn/modules/module.py”, line 1518, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
File “/home/yons/miniconda3/envs/openmmlab/lib/python3.8/site-packages/torch/nn/modules/module.py”, line 1527, in _call_impl
return forward_call(*args, **kwargs)
File “/home/yons/train/code/yolov5/models/yolo.py”, line 267, in forward
return self._forward_once(x, profile, visualize) # single-scale inference, train
File “/home/yons/train/code/yolov5/models/yolo.py”, line 167, in _forward_once
x = m(x) # run
File “/home/yons/miniconda3/envs/openmmlab/lib/python3.8/site-packages/torch/nn/modules/module.py”, line 1518, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
File “/home/yons/miniconda3/envs/openmmlab/lib/python3.8/site-packages/torch/nn/modules/module.py”, line 1527, in _call_impl
return forward_call(*args, **kwargs)
File “/home/yons/train/code/yolov5/models/common.py”, line 435, in forward
return torch.cat(x, self.d)
RuntimeError: Sizes of tensors must match except in dimension 1. Expected size 36 but got size 35 for tensor number 1 in the list.

首先打印出输入进模型的img及其shape：

发现shape都是一样的（对输入进行了处理截取固定区域），所以输入格式应该是没有问题。
然后debug到报错的concat处，发现俩shape都一致且看不出和35或36有啥关系：

找了一上午的问题，最终发现输入应该要resize成模型所需要的大小才行：

用习惯了官方写好的推理代码，其实有很多细节还是要注意的。