Skip to content
This repository has been archived by the owner on Jan 9, 2023. It is now read-only.

How can I train the network using my own datasets? #2

Open
ReedZyd opened this issue Mar 24, 2019 · 6 comments
Open

How can I train the network using my own datasets? #2

ReedZyd opened this issue Mar 24, 2019 · 6 comments

Comments

@ReedZyd
Copy link

ReedZyd commented Mar 24, 2019

I used your model file, but got an error of about 8-10 years old. I want to re-use the data set for training. What should I do?

@YOKE
Copy link

YOKE commented Mar 25, 2019

Thanks for your work! Could you share the script of train/val? @siriusdemon

@siriusdemon
Copy link
Owner

Refer to its Home Page , you can find the dataset and train/val prototxt. @ReedZyd @YOKE

I used your model file, but got an error of about 8-10 years old. I want to re-use the data set for training. What should I do?

For your problem, you can follow the origin paper: crop the face firstly, then apply the dex network to the cropped face.

@cliobeach
Copy link

do you have the code for the hole project? including face cropping and training.

@siriusdemon
Copy link
Owner

siriusdemon commented Jun 27, 2019

Nope.

For face cropping, you can use PCN or facessh or any other face detector you can find. PCN has already provided a face cropping function.

For face training, you can refer to Home Page for more details.

@cliobeach
Copy link

cliobeach commented Jul 2, 2019 via email

@cliobeach
Copy link

cliobeach commented Jul 3, 2019 via email

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

4 participants