-
Notifications
You must be signed in to change notification settings - Fork 28
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Question about reproduce 71mIOU #6
Comments
I ran into same issue. Did you solve it with the current ResNet? |
I can not reproduce the 71 mIoU claim but I got 69.7 mIoU on Cityscapes-val set which is pretty close using this for ResNet50 and this for testing Log
|
@LovPe Hi, I am not sure about the "deep_stem" param, can you explain more about the param? |
hi, guys, maybe you can try this repo: https://github.com/SegmentationBLWX/sssegmentation |
|
Thanks for Amazing code . When reproduce your mIOU encounter some problem:
1-when Achieving 71 mIOU, do you train from imageNet pretrain model?
1- Can not loading pretrain model from "models/model_store.py" because of broken link. so I use pytorch offical model instead. but some parameter can not be load due to "deep_stem" flag in ResNet model. So I set deep_stem=False. Will this hurt performance?
I use pytorch1.4 with cuda10. Currently achieve mIOU 66.1 with ResNet50 backbone
The text was updated successfully, but these errors were encountered: