反向传播时深度学习中的问题。 （蟒蛇）
[英]Issue in Deep learning while backpropagation. (Python)
问题描述
Can't seem to figure out the issue with this simple 2 layered network. 
似乎无法弄清楚这个简单的2层网络的问题。 The forward process seems to be error free, however, I am unable to figure out how to calculate the cost for w1, w2 and b1 which are the weights and bias for the first layer. 前进过程似乎没有错误，但是，我无法弄清楚如何计算w1，w2和b1的成本，它们是第一层的权重和偏差。
//forward
z1 = point[0]*w1 + point[1]*w2 + b1
z2 = sigmoid(z1)*w3 + b2
pred = sigmoid(z2)
//backward
z2_d_cost = 2 * (pred-target)
z2_d_pred = sigmoid_p(z2)
z2_cost_pred = z2_d_cost * z2_d_pred
w3 = w3 - z2*lrate*z2_cost_pred
b2 = b2 - lrate*z2_cost_pred
z1_d_pred = sigmoid_p(z1) * z2_cost_pred * w3
w1 = w1 - point[0]*lrate*z1_d_pred
w2 = w2 - point[1]*lrate*z1_d_pred
b1 = b1 - lrate*z1_d_pred
1 个解决方案
解决方案1
0 已采纳 2019-01-01 20:49:33
Nvm Figured it. Nvm想通了。 simple mistake, it should be w3 = w3 - z1*lrate*z2_cost_pred 一个简单的错误，应该是w3 = w3-z1 * lrate * z2_cost_pred
声明:本站的技术帖子网页，遵循CC BY-SA 4.0协议，如果您需要转载，请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.
相关问题
深度学习：Python中的反向传播代码 - Deep learning: the code for backpropagation in Python
Python 深度学习查找重复项 - Python Deep Learning find Duplicates
Theano错误深度学习python - Theano error deep learning python
使用Caffe进行深度学习 - Python - Deep learning using Caffe - Python
使用 python 进行深度学习（余弦相似度） - Deep learning with python (Cosine similarity)
python：神经网络和深度学习 - python: neural networks and deep learning
深度学习中的输入数据缩放问题 - input data scaling issue in Deep Learning
使用 Glove 时没有这样的文件或目录（深度学习） - No such file or directory while working with Glove (Deep learning)
在机器/深度学习上运行代码时出错 - error while running code on machine/deep learning
训练深度学习 model 时出错 - Error while training a deep learning model
相关标签