compile – Transforming Expression Graphs to Functions compile – Transforming Expression Graphs to Functions shared - defines theano.shared function - defines theano.functio...

目标函数objectives 可用的目标函数 目标函数objectives 目标函数，或称损失函数，是编译一个模型必须的两个参数之一： model . compile ( loss = 'mean_squared_error' , optimizer = 'sgd' ) 可以通过传递预定义目标函数名字指定目标函数，也可以传递一个The...

损失函数的使用 可用损失函数 mean_squared_error mean_absolute_error mean_absolute_percentage_error mean_squared_logarithmic_error squared_hinge hinge categorical_hinge logcosh categoric...

XLearning 是一款支持多种机器学习、深度学习框架的调度系统。基于Hadoop Yarn完成了对TensorFlow、MXNet、Caffe、Theano、PyTorch、Keras、XGBoost等常用框架的集成，同时具备良好的扩展性和兼容性。 架构设计 功能特性 编译&部署指南 运行示例 FAQ Authors...

一些基本概念 符号计算 张量 data_format 函数式模型 batch epochs 对新手友好的小说明 关于Python 关于深度学习 其他 一些基本概念 在开始学习Keras之前，我们希望传递一些关于Keras，关于深度学习的基本概念和技术，我们建议新手在使用Keras之前浏览一下本页面提到的内容，这将减少你学习中的困惑...

Keras: The Python Deep Learning library You have just found Keras. Multi-backend Keras and tf.keras: Guiding principles Getting started: 30 seconds to Keras Installation Confi...

nnet – Ops related to neural networks nnet – Ops related to neural networks Theano was originally developed for machine learning applications, particularlyfor the topic of deep...

shared_randomstreams – Friendly random numbers Guide Reference shared_randomstreams – Friendly random numbers Guide Since Theano uses a functional design, producing pseudo-...

Loading and Saving PyCUDA/CUDAMat/Gnumpy compatibility Understanding Memory Aliasing for Speed and Correctness Python Memory Management Multi cores support in Theano Frequentl...

Loading and Saving PyCUDA/CUDAMat/Gnumpy compatibility Understanding Memory Aliasing for Speed and Correctness Python Memory Management Multi cores support in Theano Frequentl...