转自1024深度学习
导语:本文是TensorFlow实现流行机器学习算法的教程汇集,目标是让读者可以轻松通过清晰简明的案例深入了解 TensorFlow。这些案例适合那些想要实现一些 TensorFlow 案例的初学者。本教程包含还包含笔记和带有注解的代码。
第一步:给TF新手的教程指南
1:tf初学者需要明白的入门准备
- 机器学习入门笔记:
https://github.com/aymericdamien/TensorFlow-Examples/blob/master/notebooks/0_Prerequisite/ml_introduction.ipynb
- MNIST 数据集入门笔记
https://github.com/aymericdamien/TensorFlow-Examples/blob/master/notebooks/0_Prerequisite/mnist_dataset_intro.ipynb
2:tf初学者需要了解的入门基础
- Hello World
https://github.com/aymericdamien/TensorFlow-Examples/blob/master/notebooks/1_Introduction/helloworld.ipynb
https://github.com/aymericdamien/TensorFlow-Examples/blob/master/examples/1_Introduction/helloworld.py
- 基本操作
https://github.com/aymericdamien/TensorFlow-Examples/blob/master/notebooks/1_Introduction/basic_operations.ipynb
https://github.com/aymericdamien/TensorFlow-Examples/blob/master/examples/1_Introduction/basic_operations.py
3:tf初学者需要掌握的基本模型
- 最近邻:
https://github.com/aymericdamien/TensorFlow-Examples/blob/master/notebooks/2_BasicModels/nearest_neighbor.ipynb
https://github.com/aymericdamien/TensorFlow-Examples/blob/master/examples/2_BasicModels/nearest_neighbor.py
- 线性回归:
https://github.com/aymericdamien/TensorFlow-Examples/blob/master/notebooks/2_BasicModels/linear_regression.ipynb
https://github.com/aymericdamien/TensorFlow-Examples/blob/master/examples/2_BasicModels/linear_regression.py
- Logistic 回归:
https://github.com/aymericdamien/TensorFlow-Examples/blob/master/notebooks/2_BasicModels/logistic_regression.ipynb
https://github.com/aymericdamien/TensorFlow-Examples/blob/master/examples/2_BasicModels/logistic_regression.py
4:tf初学者需要尝试的神经网络
- 多层感知器:
https://github.com/aymericdamien/TensorFlow-Examples/blob/master/notebooks/3_NeuralNetworks/multilayer_perceptron.ipynb
https://github.com/aymericdamien/TensorFlow-Examples/blob/master/examples/3_NeuralNetworks/multilayer_perceptron.py
- 卷积神经网络:
https://github.com/aymericdamien/TensorFlow-Examples/blob/master/notebooks/3_NeuralNetworks/convolutional_network.ipynb
https://github.com/aymericdamien/TensorFlow-Examples/blob/master/examples/3_NeuralNetworks/convolutional_network.py
- 循环神经网络(LSTM):
https://github.com/aymericdamien/TensorFlow-Examples/blob/master/notebooks/3_NeuralNetworks/recurrent_network.ipynb
https://github.com/aymericdamien/TensorFlow-Examples/blob/master/examples/3_NeuralNetworks/recurrent_network.py
- 双向循环神经网络(LSTM):
https://github.com/aymericdamien/TensorFlow-Examples/blob/master/notebooks/3_NeuralNetworks/bidirectional_rnn.ipynb
https://github.com/aymericdamien/TensorFlow-Examples/blob/master/examples/3_NeuralNetworks/bidirectional_rnn.py
- 动态循环神经网络(LSTM)
https://github.com/aymericdamien/TensorFlow-Examples/blob/master/examples/3_NeuralNetworks/dynamic_rnn.py
- 自编码器
https://github.com/aymericdamien/TensorFlow-Examples/blob/master/notebooks/3_NeuralNetworks/autoencoder.ipynb
https://github.com/aymericdamien/TensorFlow-Examples/blob/master/examples/3_NeuralNetworks/autoencoder.py
5:tf初学者需要精通的实用技术
- 保存和恢复模型
https://github.com/aymericdamien/TensorFlow-Examples/blob/master/notebooks/4_Utils/save_restore_model.ipynb
https://github.com/aymericdamien/TensorFlow-Examples/blob/master/examples/4_Utils/save_restore_model.py
- 图和损失可视化
https://github.com/aymericdamien/TensorFlow-Examples/blob/master/notebooks/4_Utils/tensorboard_basic.ipynb
https://github.com/aymericdamien/TensorFlow-Examples/blob/master/examples/4_Utils/tensorboard_basic.py
- Tensorboard——高级可视化
https://github.com/aymericdamien/TensorFlow-Examples/blob/master/examples/4_Utils/tensorboard_advanced.py
5:tf初学者需要的懂得的多GPU基本操作
- 多 GPU 上的基本操作
https://github.com/aymericdamien/TensorFlow-Examples/blob/master/notebooks/5_MultiGPU/multigpu_basics.ipynb
https://github.com/aymericdamien/TensorFlow-Examples/blob/master/examples/5_MultiGPU/multigpu_basics.py
6:案例需要的数据集
有一些案例需要 MNIST 数据集进行训练和测试。运行这些案例时,该数据集会被自动下载下来(使用 input_data.py)。
MNIST数据集笔记:https://github.com/aymericdamien/TensorFlow-Examples/blob/master/notebooks/0_Prerequisite/mnist_dataset_intro.ipynb
官方网站:http://yann.lecun.com/exdb/mnist/
第二步:为TF新手准备的各个类型的案例、模型和数据集
初步了解:TFLearn TensorFlow
接下来的示例来自TFLearn,这是一个为 TensorFlow 提供了简化的接口的库。里面有很多示例和预构建的运算和层。
使用教程:TFLearn 快速入门。通过一个具体的机器学习任务学习 TFLearn 基础。开发和训练一个深度神经网络分类器。
TFLearn地址:https://github.com/tflearn/tflearn
示例:https://github.com/tflearn/tflearn/tree/master/examples
预构建的运算和层:http://tflearn.org/doc_index/#api
笔记:https://github.com/tflearn/tflearn/blob/master/tutorials/intro/quickstart.md
基础模型以及数据集
- 线性回归,使用 TFLearn 实现线性回归
https://github.com/tflearn/tflearn/blob/master/examples/basics/linear_regression.py
- 逻辑运算符。使用 TFLearn 实现逻辑运算符
https://github.com/tflearn/tflearn/blob/master/examples/basics/logical.py
- 权重保持。保存和还原一个模型
https://github.com/tflearn/tflearn/blob/master/examples/basics/weights_persistence.py
- 微调。在一个新任务上微调一个预训练的模型
https://github.com/tflearn/tflearn/blob/master/examples/basics/finetuning.py
- 使用 HDF5。使用 HDF5 处理大型数据集
https://github.com/tflearn/tflearn/blob/master/examples/basics/use_hdf5.py
- 使用 DASK。使用 DASK 处理大型数据集
https://github.com/tflearn/tflearn/blob/master/examples/basics/use_dask.py
计算机视觉模型及数据集
- 多层感知器。一种用于 MNIST 分类任务的多层感知实现
https://github.com/tflearn/tflearn/blob/master/examples/images/dnn.py
- 卷积网络(MNIST)。用于分类 MNIST 数据集的一种卷积神经网络实现
https://github.com/tflearn/tflearn/blob/master/examples/images/convnet_mnist.py
- 卷积网络(CIFAR-10)。用于分类 CIFAR-10 数据集的一种卷积神经网络实现
https://github.com/tflearn/tflearn/blob/master/examples/images/convnet_cifar10.py
- 网络中的网络。用于分类 CIFAR-10 数据集的 Network in Network 实现
https://github.com/tflearn/tflearn/blob/master/examples/images/network_in_network.py
- Alexnet。将 Alexnet 应用于 Oxford Flowers 17 分类任务
https://github.com/tflearn/tflearn/blob/master/examples/images/alexnet.py
- VGGNet。将 VGGNet 应用于 Oxford Flowers 17 分类任务
https://github.com/tflearn/tflearn/blob/master/examples/images/vgg_network.py
- VGGNet Finetuning (Fast Training)。使用一个预训练的 VGG 网络并将其约束到你自己的数据上,以便实现快速训练
https://github.com/tflearn/tflearn/blob/master/examples/images/vgg_network_finetuning.py
- RNN Pixels。使用 RNN(在像素的序列上)分类图像
https://github.com/tflearn/tflearn/blob/master/examples/images/rnn_pixels.py
- Highway Network。用于分类 MNIST 数据集的 Highway Network 实现
https://github.com/tflearn/tflearn/blob/master/examples/images/highway_dnn.py
- Highway Convolutional Network。用于分类 MNIST 数据集的 Highway Convolutional Network 实现
https://github.com/tflearn/tflearn/blob/master/examples/images/convnet_highway_mnist.py
- Residual Network (MNIST) 。应用于 MNIST 分类任务的一种瓶颈残差网络(bottleneck residual network)
https://github.com/tflearn/tflearn/blob/master/examples/images/residual_network_mnist.py
- Residual Network (CIFAR-10)。应用于 CIFAR-10 分类任务的一种残差网络
https://github.com/tflearn/tflearn/blob/master/examples/images/residual_network_cifar10.py
- Google Inception(v3)。应用于 Oxford Flowers 17 分类任务的谷歌 Inception v3 网络
https://github.com/tflearn/tflearn/blob/master/examples/images/googlenet.py
- 自编码器。用于 MNIST 手写数字的自编码器
https://github.com/tflearn/tflearn/blob/master/examples/images/autoencoder.py
自然语言处理模型及数据集
- 循环神经网络(LSTM),应用 LSTM 到 IMDB 情感数据集分类任
https://github.com/tflearn/tflearn/blob/master/examples/nlp/lstm.py
- 双向 RNN(LSTM),将一个双向 LSTM 应用到 IMDB 情感数据集分类任务:
https://github.com/tflearn/tflearn/blob/master/examples/nlp/bidirectional_lstm.py
- 动态 RNN(LSTM),利用动态 LSTM 从 IMDB 数据集分类可变长度文本:
https://github.com/tflearn/tflearn/blob/master/examples/nlp/dynamic_lstm.py
- 城市名称生成,使用 LSTM 网络生成新的美国城市名:
https://github.com/tflearn/tflearn/blob/master/examples/nlp/lstm_generator_cityname.py
- 莎士比亚手稿生成,使用 LSTM 网络生成新的莎士比亚手稿:
https://github.com/tflearn/tflearn/blob/master/examples/nlp/lstm_generator_shakespeare.py
- Seq2seq,seq2seq 循环网络的教学示例:
https://github.com/tflearn/tflearn/blob/master/examples/nlp/seq2seq_example.py
- CNN Seq,应用一个 1-D 卷积网络从 IMDB 情感数据集中分类词序列
https://github.com/tflearn/tflearn/blob/master/examples/nlp/cnn_sentence_classification.py
强化学习案例
- Atari Pacman 1-step Q-Learning,使用 1-step Q-learning 教一台机器玩 Atari 游戏:
https://github.com/tflearn/tflearn/blob/master/examples/reinforcement_learning/atari_1step_qlearning.py
第三步:为TF新手准备的其他方面内容
- Recommender-Wide&Deep Network,推荐系统中 wide & deep 网络的教学示例:
https://github.com/tflearn/tflearn/blob/master/examples/others/recommender_wide_and_deep.py
- Spiral Classification Problem,对斯坦福 CS231n spiral 分类难题的 TFLearn 实现:
https://github.com/tflearn/tflearn/blob/master/examples/notebooks/spiral.ipynb
- 层,与 TensorFlow 一起使用 TFLearn 层:
https://github.com/tflearn/tflearn/blob/master/examples/extending_tensorflow/layers.py
- 训练器,使用 TFLearn 训练器类训练任何 TensorFlow 图:
https://github.com/tflearn/tflearn/blob/master/examples/extending_tensorflow/layers.py
- Bulit-in Ops,连同 TensorFlow 使用 TFLearn built-in 操作:
https://github.com/tflearn/tflearn/blob/master/examples/extending_tensorflow/builtin_ops.py
- Summaries,连同 TensorFlow 使用 TFLearn summarizers:
https://github.com/tflearn/tflearn/blob/master/examples/extending_tensorflow/summaries.py
- Variables,连同 TensorFlow 使用 TFLearn Variables:
https://github.com/tflearn/tflearn/blob/master/examples/extending_tensorflow/variables.py