keras-YOLOv3-mobilenet icon indicating copy to clipboard operation
keras-YOLOv3-mobilenet copied to clipboard

How you used MobileNet model?

Open meet-minimalist opened this issue 5 years ago • 5 comments

Have you used the pretrained MobileNet model and construct yolo layers on top? And used pretrained MobileNet weights for Mobilenet initialization?

Or you just took the architecture of MobileNet and add yolo layers and started training from scratch using voc or coco?

meet-minimalist avatar Apr 10 '19 10:04 meet-minimalist

I used MobileNet weights for MobileNet initialization, and feed VOC2007 datasets into network. But, i got train loss non-convergence(train loss : 30, val loss:26),whaterver i decrease learning rate, i can't got train loss convergence.

ZH-Lee avatar May 18 '19 04:05 ZH-Lee

@ZH-Lee so what you did to make it converge?

meet-minimalist avatar May 18 '19 05:05 meet-minimalist

@ZH-Lee so what you did to make it converge?

i didn't make it converge. maybe the datasets is small??? I haven't solve the problem.

ZH-Lee avatar May 20 '19 02:05 ZH-Lee

@ZH-Lee ohk. 👍

meet-minimalist avatar May 20 '19 04:05 meet-minimalist

I used MobileNet weights for MobileNet initialization, and feed VOC2007 datasets into network. But, i got train loss non-convergence(train loss : 30, val loss:26),whaterver i decrease learning rate, i can't got train loss convergence.

How do you train with VOC dataset? just transfer voc format to coco in this code? i attemp to do it but the map is much low than yolov2 on voc. Why ?

ruoruo6 avatar May 23 '19 01:05 ruoruo6