Result not showing in the output folder

[IrvingF7]

Hi!
I am trying to run the classifier based on YOLOv3, and after I run it with yolov3.weights the outputs are just the original pics. If I run it with yolov3.pt it would show error msg like this
size mismatch for module_list.105.Conv2d.bias: copying a param with shape torch.Size([255]) from checkpoint, the shape in current model is torch.Size([21]).

I am thinking that the trained model that I am using is not the intended model for gender detection.

BTW I am not using docker as docker build command always fails. Do you think that may be the problem?

Thanks!

[Pari-singh]

Hi @IrvingF7, thanks for your interest in my work.

I am not sure where you are taking the weights yolov3.pt from, but I suspect it is a yolo detector most likely trained on COCO classes, which doesn't have man vs woman. It has one class for Person, including both the genders.

Let me know if any further doubts.

[IrvingF7]

Hi!
Thank you for your reply! I apologize for the late reply. For some reasons Github is not sending me notification.

I am getting yolov3.pt and yolov3.weights from Releases of this repo under the tag name "weights".
After reading your description in Yolov3_gender_detection I was under the impression that there is a trained detector ready to detect men vs women so I looked around and was only able to find that one.

If possible could you provide a trained model that can run on the sample images to detect men and women? Really appreciate it!