search

guoqiangqi / PFLD

Implementation of PFLD A Practical Facial Landmark Detector , reference to https://arxiv.org/pdf/1902.10859.pdf
626 stars 166 forks source link

can you explain how to generate training dataset? #4

Open alexliyang opened 5 years ago

alexliyang commented 5 years ago

thank you for opening your code. I croped some faces , and how to generate the training dataset?

guoqiangqi commented 5 years ago

thank you for opening your code. I croped some faces , and how to generate the training dataset?

Have you got the face landmark annotation file?

alexliyang commented 5 years ago

I have not got the annotation files. which dataset do you use? and do you have the tool to label samples to generate the annotation file?

guoqiangqi commented 5 years ago

I have trained model with WFLW dataset，it`s a hard work for you to label the face images,there are many open dataset you can download ,such as 300-W（68 points）and WFLW(98 points)

john-bao-git commented 4 years ago

How do I go about training for a 300W model? Is it possible to release a pre-trained model?

I was going through your sh file and it was mentioned that I needed:

save_model=./models2/model4 pre_model=./models2/model3

However, I didn't find the models2 directories to start training.

Thank you for your time.

AIwzx commented 4 years ago

where the pre_model

github-luffy commented 4 years ago

You can try PFLD_68points_Pytorch.