Slow convergence
xmjiayou opened this issue · 2 comments
xmjiayou commented
Slow convergence when training the network.I tarin the model using remote sensing images(segmented ;num:160000）on single GTX1080TI . it have been worked more than one week for edge model,but the precision could not be over 0.5.
Is wrong with my dataset?or other problem?
youyuge34 commented
@xmjiayou There must be something wrong, transfer learning is quite fast that u can see the results obviously better in just one day!
xmjiayou commented
ok,thanks! I will check it more carefully to find the problem.