BVLC / caffe

Caffe: a fast open framework for deep learning.
http://caffe.berkeleyvision.org/
Other
34.06k stars 18.7k forks source link

Solver in caffe #6914

Open jcompi opened 4 years ago

jcompi commented 4 years ago

Hello, somebody can explain me how to design the hyper parameters in the solver file, i get 592 train cases and 176 test cases, train/test batch = 10 My model is very simple, one FC to input, two hidden FC layer and finally an accuracy layer and softmax loss and i get only two classes.