Yolov5 advanced 8 format conversion between high and low versions

yolov5 6.0 after pt The weight file has changed , Cannot be on 5.0 pt A mixture of , Lead to the original 5.0 The version often needs to train the data set by itself .
Here is a yolov5 5.0 call newest yolov5 6.0 6.1 Method of weight file .
Directly applicable 6.0 Data sets , perform

python detect.py --source 0

AttributeError: Can’t get attribute ‘SPPF’ on <module ‘models.common’ from ‘D:\PycharmProjects\deepsort3.0\Yolov5_StrongSORT_OSNet-3.0\yolov5\models\common.py’>
Is due to 5.0 No, SPPF class , take 6.0 Of SPPF Class to model Under the common.py And add it to the head import warnings
Continue to execute the above command , Tips
if ‘youtube.com/’ in url or ‘youtu.be/’ in url: # if source is YouTube video
TypeError: argument of type ‘int’ is not iterable
This problem is explained by advanced six , Only need dataset.py in url Turn into str(url), Two consecutive places .
Re execution ,
RuntimeError: The size of tensor a (80) must match the size of tensor b (56) at non-singleton dimension 3
This is the hardest thing to do , This is the function of the new version 2 Dimensional change 3 Vivian , open model Under the yolo.py, Prepare for 6.0 Of yolo.py Partial assignment , Two functions def forward(self, x): and def _make_grid(self, nx=20, ny=20, i&#