search

zhangda1018 / yolov5-crowdhuman

Train yolov5 on crowdhuman dataset.
GNU General Public License v3.0
28 stars 6 forks source link

Achieved mAP? #1

Closed mikel-brostrom closed 2 years ago

mikel-brostrom commented 2 years ago

What was the achieved mAP for your model?

zhangda1018 commented 2 years ago

image I still train my model now, and this is what I've attained now.

mikel-brostrom commented 2 years ago

Thanks for sharing. I trained a model, starting with a Yolov5m pretrained on COCO, 100 epochs, imgsz 640, only people no heads, with standard setup. My result: metrics/mAP_0.5:0.95: 0.45613. Yolov5x gives me much better results: 0.66814. I am 70 epoch in, imgz 1280.

zhangda1018 commented 2 years ago

Thanks for sharing. I trained a model, starting with a Yolov5m pretrained on COCO, 100 epochs, imgsz 640, only people no heads, with standard setup. My result: metrics/mAP_0.5:0.95: 0.45613. Yolov5x gives me much better results: 0.66814. I am 70 epoch in, imgz 1280.

Thanks, I will have a try. By the way,do you apply the result to yolov5-deepsort, how is it?

mikel-brostrom commented 2 years ago

Will try the larger model by the end of this week :)

zhangda1018 commented 2 years ago

Will try the larger model by the end of this week :)

Thanks a lot. Hope your result!

zhangda1018 commented 2 years ago

Will try the larger model by the end of this week :)

Hello，Can I ask your result？

mikel-brostrom commented 2 years ago

mAP_0.5:0.95 | 0.69388 Yolov5x, imgsz: 1280