qiexing / face-landmark-localization

cnn network predict face landmarks (68 points) and head pose (3d pose, yaw,roll,pitch).
410 stars 184 forks source link

train_val.prototxt and solver.prototxt #11

Open GabrielLin opened 8 years ago

GabrielLin commented 8 years ago

Dear qiexing, Could you please provide the train_val.prototxt and solver.prototxt of Caffe? Many thanks.

qiexing commented 8 years ago

If you are interested in how to use caffe to regress landmark, you can see this: https://github.com/qiexing/caffe-regression first.

GabrielLin commented 8 years ago

Thank you for your reply. That project is useful for learning how to use regression of Caffe. The reason why I want train_val.prototxt and solver.prototxt in this project is that I want to train the model by myself. Do you convenience to share them?

qiexing commented 8 years ago

I will try to find the code, but not sure i can find it.

GabrielLin commented 8 years ago

Thanks. I strongly hope that you can find it.

GabrielLin commented 8 years ago

Dear qiexing, have you find the prototxt files? Thanks.

qiexing commented 8 years ago

@GabrielLin add train_val.prototxt and solver

GabrielLin commented 8 years ago

Thank you very much. You are a really a person.

Sun1992 commented 8 years ago

Hi qiexing, I like your model. What's your training dataset of 68 points? Is it the same as the dlib?