Notes, Codes, and Tutorials for the Deep Learning Course at ChinaHadoop
注意每一份代码分别有Jupyter Notebook, Python, 以及HTML三种形式,大家可以按照自己的需求阅读,学习或运行。 运行时需要注意anaconda的版本问题,anaconda2-5.0.0与anaconda3-5.0.0分别对应python2.7与python3.6环境。
重要参考资料:
学习资料:
- Effective TensorFlow - TensorFlow tutorials and best practices.
- Finch - Many Machine Intelligence models implemented (mainly tensorflow, sometimes pytorch / mxnet)
- Pytorch Tutorials - PyTorch Tutorial for Deep Learning Researchers.
- MXNet the straight dope - An interactive book on deep learning. Much easy, so MXNet. Wow.
代码示例:TensorFlow基础与线性回归模型(TensorFlow, PyTorch)
- MNIST数据集演示
- TensorFlow基础
- 线性回归模型-TensorFlow
- 线性回归模型-PyTorch
- 线性回归模型-MXNet (contributed by LinkHS)
代码示例:K近邻算法,线性分类,以及多层神经网络(TensorFlow, PyTorch)
代码示例:卷积神经网络的基础实现(TensorFlow)
代码示例:卷积神经网络的进阶实现(TensorFlow)
代码示例:深度神经网络-图像识别与分类(TensorFlow, PyTorch)
- 安装TensorLayer (中文文档参见此处,此后复杂实现均推荐使用TensorLayer高级API库,同时可以结合TF-Slim与Keras)
pip install git+https://github.com/zsdonghao/tensorlayer.git
- 安装OpenCV python接口
conda install -c menpo opencv3
- 所需数据集下载:
data.zip
: [微云][百度云] (覆盖./05_Image_recognition_and_classification/data
文件夹) - 所需模型下载:
vgg19.npz
[微云][百度云] (放置于./05_Image_recognition_and_classification
文件夹下) - 所需模型下载:
inception_v3.ckpt
[微云][百度云] (放置于./05_Image_recognition_and_classification
文件夹下)
- Class Activation Mapping (CAM)示例 (完整实现可参考此处)
代码示例:目标检测模型示例 (TensorFlow, PyTorch)
- 所需模型下载:
ssd_mobilenet_v1_coco_11_06_2017
: [微云] (解压并置于06_Object_detection/Object_Detection_Tensorflow_API_demo/object_detection/
文件夹下)
-
[
SSD: Single Shot Multibox Detector
] (TensorFlow实现, PyTorch实现) -
[
YOLO
,YOLOv2
] (TensorFlow实现, PyTorch实现)