search

WongKinYiu / YOLO

An MIT rewrite of YOLOv9
MIT License
597 stars 64 forks source link

predict on image with custom trained model #71

Open ck-amrahd opened 2 months ago

ck-amrahd commented 2 months ago

Issue Description I have trained the model on the custom dataset, tried to do inference using the inference notebook, created a model using create_model, passed the class_num to be my custom number, and tried to load the trained model using load state dict and torch load, but it's not working as expected. Do you have any examples of this?

Additional Context Add any other context or screenshots about the issue here.

Future Considerations Please suggest any potential future improvements related to this issue.

yjmm10 commented 2 months ago

Did you have the value of AP during the validation phase of training.

I got the value that approach 0

ck-amrahd commented 2 months ago

Nope, I didn't get anything. It saved the trained weights, that's all.