Skip to content
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

performance difference between fmobilenet and fmobilenetv2 #34

Closed
zhenglaizhang opened this issue Feb 7, 2018 · 4 comments
Closed

performance difference between fmobilenet and fmobilenetv2 #34

zhenglaizhang opened this issue Feb 7, 2018 · 4 comments

Comments

@zhenglaizhang
Copy link

Hi, from the paper and readme, it seems the mobilenet m1 == fmobilenet ==LMobileNetE , and curiors the performance of fmobilenetv2.

@nttstar
Copy link
Collaborator

nttstar commented Feb 7, 2018

I haven't finished a whole training process with fmobilenetv2 yet, you can try it.

@zhenglaizhang
Copy link
Author

@nttstar Yeah, I will do in the near future, I have just tried to train fmobilenet, but find the checkpoint params file is around 227M, it is so big, trained with following configurations (I used the memonger so that I could set the per batch size as 128, for the pool 1060 6G memory -:( )

--network m1 --loss-type 4 --margin-m 0.5 --per-batch-size 128 --memonger ...

insightface paper states the size of mobilenet is around 112 Mb, I am a bit confused here.

@nttstar
Copy link
Collaborator

nttstar commented Feb 8, 2018

remove last weight matrix by deploy/model_slim.py

@zhenglaizhang
Copy link
Author

@nttstar Oh, thanks man, I should have thought of it -:)

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

No branches or pull requests

2 participants