官术网_书友最值得收藏!

Setting up your workspace

Before we can start, you will need to set up your workspace. The examples in this book are all meant to run in a Jupyter notebook. Jupyter notebooks are an interactive development environment mostly used for data-science applications and are considered the go-to environment to build data-driven applications in.

You can run Jupyter notebooks either on your local machine, on a server in the cloud, or on a website such as Kaggle.

Note

Note: All code examples for this book can be found here: https://github.com/PacktPublishing/Machine-Learning-for-Finance and for chapter 1 refer the following link: https://www.kaggle.com/jannesklaas/machine-learning-for-finance-chapter-1-code.

Deep learning is computer intensive, and the data used in the examples throughout this book are frequently over a gigabyte in size. It can be accelerated by the use of Graphics Processing Units (GPUs), which were invented for rendering video and games. If you have a GPU enabled computer, you can run the examples locally. If you do not have such a machine, it is recommended to use a service such as Kaggle kernels.

Learning deep learning used to be an expensive endeavor because GPUs are an expensive piece of hardware. While there are cheaper options available, a powerful GPU can cost up to $10,000 if you buy it and about $0.80 an hour to rent it in the cloud.

If you have many, long-running training jobs, it might be worth considering building a "deep learning" box, a desktop computer with a GPU. There are countless tutorials for this online and a decent box can be assembled for as little as a few hundred dollars all the way to $5,000.

The examples in this book can all be run on Kaggle for free, though. In fact, they have been developed using this site.

主站蜘蛛池模板: 连州市| 平陆县| 阿勒泰市| 陈巴尔虎旗| 咸宁市| 祁门县| 呼和浩特市| 科技| 长宁区| 天峨县| 仪征市| 宜川县| 大宁县| 三原县| 凤台县| 广安市| 南通市| 石屏县| 合肥市| 纳雍县| 三原县| 本溪| 金门县| 桂东县| 佛坪县| 岳普湖县| 丰镇市| 措勤县| 宜川县| 博乐市| 怀来县| 兰坪| 玛纳斯县| 峨山| 中卫市| 灵寿县| 社会| 兴海县| 拉萨市| 施秉县| 高安市|