Skip to content
This repository has been archived by the owner on May 28, 2021. It is now read-only.

Implementing different backbone to this YOLOv3 #48

Open
hilmandayo opened this issue Sep 7, 2019 · 1 comment
Open

Implementing different backbone to this YOLOv3 #48

hilmandayo opened this issue Sep 7, 2019 · 1 comment

Comments

@hilmandayo
Copy link

hilmandayo commented Sep 7, 2019

Hello.

First of all, thank you for this awesome repo!

I have one question.
I want to change the backbone from darknet to something else.
After inspected and tested the code, I found out that the things that I need to do are (roughly):

  1. Make sure that the new backbone net output three tensors from different layers in forward method
  2. While referring to nets/backbone/darknet.py, add some needed implementation details to the new backbone (like attributes and methods)
  3. Set the new backbone in params.py

Is this correct? Or is there any other things that I need to do or be careful of?
Thanks.

@hilmandayo hilmandayo changed the title Implementing different backend to this YOLOv3 Implementing different backbone to this YOLOv3 Sep 7, 2019
@mei123hao
Copy link

Hello.

First of all, thank you for this awesome repo!

I have one question.
I want to change the backbone from darknet to something else.
After inspected and tested the code, I found out that the things that I need to do are (roughly):

  1. Make sure that the new backbone net output three tensors from different layers in forward method
  2. While referring to nets/backbone/darknet.py, add some needed implementation details to the new backbone (like attributes and methods)
  3. Set the new backbone in params.py

Is this correct? Or is there any other things that I need to do or be careful of?
Thanks.
Hi, I try to change the backbone same as you. However, the training loss reduce to 0.3 and stop falling, I try to test the saved weight file but got nothing. Do you solve your problems or have any adiveces?

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants