save and load nueral network parameters when using python lasagne
import cPickle as pickle
sys.setrecursionlimit(10000) # you may need this if the network is large
with open("model_file", 'wb') as f:
pickle.dump(nolearnnet , f, -1)但是,如果据说在GPU上训练的神经网络要用读取出来用CPU做计
翻译
2017-04-04 15:53:42 ·
431 阅读 ·
0 评论