olddocks / caffe-facialkp Goto Github PK

Hey Prabhu,

I’ve been able to run your code. Thanks for putting it out there. Caffe really lacks a good regression example.

Do you know why all of my prediction results for different images are the same?

I appreciate your answer :)

facialkp.prototxt has 2 relu1 relu4 layers.

data4DL = np.zeros([BATCH_SIZE,1,1,1]) # need to create 4D array as output, first value is batch_size, last number of outputs

what is data4DL in output.py? The number of outputs is 30. why the last value is set as 1? shouldnt it be 30?

Hi,

I am working somehow on the same subject as this repository. I have some flow images of a some scenes and want to train caffe to do regression using eulidean loss to predict this gray scale flow map. could you instruct me how to make the data layer using hdf5 and how to configure accuracy layer?

thanks