search

michalfaber / keras_Realtime_Multi-Person_Pose_Estimation

Keras version of Realtime Multi-Person Pose Estimation project
Other
780 stars 372 forks source link

If the datasets have no miss mask and all mask, then how I should deal with miss mask and all mask? #4

Open xiangweifeng opened 7 years ago

xiangweifeng commented 7 years ago

Let them all '1'?

michalfaber commented 7 years ago

miss mask and all mask are required only for training - I work on that

Marcovaldong commented 7 years ago

I want to transfer this model to a new dataset which has 14 keypoints for one person, I need to train the model on the new dataset, so how I should deal with miss mask and all mask