本文整理汇总了Python中mlp.MLP.get_gradient方法的典型用法代码示例。如果您正苦于以下问题：Python MLP.get_gradient方法的具体用法？Python MLP.get_gradient怎么用？Python MLP.get_gradient使用的例子？那么恭喜您, 这里精选的方法代码示例或许可以为您提供帮助。您也可以进一步了解该方法所在类mlp.MLP的用法示例。
在下文中一共展示了MLP.get_gradient方法的2个代码示例，这些例子默认根据受欢迎程度排序。您可以为喜欢或者感觉有用的代码点赞，您的评价将有助于系统推荐出更棒的Python代码示例。
示例1: open
# 需要导入模块: from mlp import MLP [as 别名]
# 或者: from mlp.MLP import get_gradient [as 别名]
with open('debug_nnet.pickle') as f:
init_param = pickle.load(f)
init_param = numpy.concatenate([i.flatten() for i in init_param])
mlp.packParam(init_param)
with open('debug_data.pickle') as f:
data = pickle.load(f)
X = data[0]
Y = data[1]
with open('HJv.pickle') as f:
HJv_theano = pickle.load(f)
num_param = numpy.sum(mlp.sizes)
batch_size = 100
grad,train_nll,train_error=mlp.get_gradient(X,Y,batch_size)
d = 1.0*numpy.ones((num_param,))
col = mlp.get_Gv(X, Y, batch_size, d)
#print 'Some col:'
#print col
"""
grad,train_nll,train_error=mlp.get_gradient(X,Y,2)
v=numpy.zeros(num_param)
mlp.forward(X)
O = mlp.layers[-1].output
S = mlp.layers[-1].linear_output
#nll.append(mlp.Cost(Y))示例2:
# 需要导入模块: from mlp import MLP [as 别名]
# 或者: from mlp.MLP import get_gradient [as 别名]
numpy.random.seed(18877)
numpy.random.shuffle(train_cg_X)
numpy.random.seed(18877)
numpy.random.shuffle(train_cg_Y)
train_cg_X_cur = train_cg_X[cg_chunk_index*cg_chunk_size:(cg_chunk_index+1)*cg_chunk_size,:]
train_cg_Y_cur = train_cg_Y[cg_chunk_index*cg_chunk_size:(cg_chunk_index+1)*cg_chunk_size]
cg_chunk_index = cg_chunk_index+1
nll=[]
error=[]
print "Iter: %d ..."%(i), "Lambda: %f"%(mlp._lambda)
grad,train_nll,train_error = mlp.get_gradient(train_gradient_X, train_gradient_Y, batch_size)
delta, next_init, after_cost = mlp.cg(-grad, train_cg_X_cur, train_cg_Y_cur, batch_size, next_init, 1)
Gv = mlp.get_Gv(train_cg_X_cur,train_cg_Y_cur,batch_size,delta)
delta_cost = numpy.dot(delta,grad+0.5*Gv)
before_cost = mlp.quick_cost(numpy.zeros((num_param,)), train_cg_X_cur, train_cg_Y_cur, batch_size)
l2norm = numpy.linalg.norm(Gv + mlp._lambda*delta + grad)
print "Residual Norm: ",l2norm
print 'Before cost: %f, After cost: %f'%(before_cost,after_cost)
param = mlp.flatParam() + delta