-
Notifications
You must be signed in to change notification settings - Fork 32
i have an issue #6
Comments
I think you use the wrong checkpoint file format. Can you send me that .pt file ? |
I use YOLO V5 m net , it works well on other codes |
If you are using the trained model from ultralytics, you will have to get the state dict of the model and save it as the .pt file (meaning not to use the author's save function). To do this, I suggest looking into the train.py from author's code, search for the line where the model loading the state dict (something like Then, the |
Hi @kaylode Conversion scriptI bodged the following script, "It has parts of train.py from ultralytics, it exported a file 'model.pt'
Running vehicle countingI ran
|
You are correct @nishontan , but with little modification that I forgot to mention: From
To
Unfortunately, as I have tested recently, original YoloV5 repo has updated their model to newer version. So that if you train your model with this new version , it may not be loaded correctly in this repo. I'm trying to figure out what they've updated and I will update mine accordingly. |
Thanks! It exported, It ran in your pipeline but it did not detect. Here's the script
|
Hey @nishontan , |
It did not work for me.
The weights I used were trained a month back. |
Can't get attribute 'Model' on <module 'models.yolo' from 'E:\Python\vehicle-counting-master\vehicle-counting-master\models\yolo\init.py'>
i run
python run.py --input_path='moos.mp4' --output_path="results" --weight="best.pt"
The text was updated successfully, but these errors were encountered: