Open dyther opened 5 years ago
Have you done with training part with theano?
Have you done with training part with theano?
Yes. But I did some little change to run it
and ,this is my result when I trained on subset0
@dyther can I know what changes you did? Apart from that you can save the model and use model.predict() from keras. Use keras as frontend and tensorflow as backend it will work. About results you can create the numpy array for corresponding test images check with its corresponding marked annotations.
weight = cp.load(open('filters.pkl'))
What's the filters.pkl?