Programmer Think

Author Zhao Luyang In 2016, Chen Tianqi's team proposed "gradient/activation checkpointing" and other technologies related to sub linear memory optimization [1], aiming to reduce the memory occupation caused by intermediate activation in the process of deep learning and training. Checkpointing technology is a kind of sub linear mem ...

Interested children, especially children's shoes, especially in the application of big data and artificial intelligence in quantification, can pay attention to my official account. If the script code is not clear, please see the official account number: datahomex: Through the introduction of the previous two articles in this series, we must ...

Source: https://iyinst.github.io/2021/05/19/%E6%9C%BA%E5%99%A8%E5%AD%A6%E4%B9%A0%E7%AC%94%E8%AE%B0-%E6%89%8B%E6%92%95GBDT/ The principle of gradient lifting decision tree (GBDT) will not be repeated. Unfamiliar students can refer to Mr. Li Hang's statistical learning method. Here is an implementation of GBDT. Before coding, we should first co ...

1. Polynomial Regression If your data points are obviously unsuitable for linear regression (straight lines between data points), polynomial regression may be an ideal choice. Like linear regression, polynomial regression uses the relationship between variables x and y to find the best way to plot data point lines. 2. Working principle ...

🌟 Last time we talked about the decision tree algorithm, which is based on sklearn. This time, we want to learn about naive Bayes, what is "naive". The meaning of simplicity is that the features are independent of each other and have no correlation. Small partners interested in previous content can refer to the following content 👇: ...

Time series analysis ARIMA Stationarity: stationarity requires that the fitting curve obtained through the sample time series can continue along the existing form "inertia" in the future Stationarity requires that the mean and variance of the series do not change significantly Strict and weak stability: Yan pingwen: the distr ...

[translated from: Gradient Descent With RMSProp from Scratch] [Note: Jason Brownlee likes PhD's articles very much, so he will do some translation and learning practice in his spare time. Here is the practice record of the corresponding work, hoping to help people in need!] Gradient descent is an optimization algorithm, which follows the neg ...

To illustrate the environment, the code is implemented in C + +, but the matrix operation is not implemented in C + +. Making wheels by yourself is a waste of time and unnecessary. Therefore, the Eigen library is used for matrix operation, and the codes of other functions are implemented by yourself. 1, Environment configuration The Eigen mat ...

The basic communication mechanism in ROS mainly has the following three implementation strategies: Topic communication (publish subscribe mode) Service communication (request response mode) Parameter server (parameter sharing mode) 1. Topic communication: publish subscribe mode One node publishes messages to the topic, and the other node ...

Clustering algorithm - Kmeans 1. Unsupervised learning and clustering algorithm Supervised learning model algorithm, the sample data required by the model needs both characteristic matrix X and real label y.Unsupervised learning means that the model only needs to use the characteristic matrix X without the real label y. clustering algorithm i ...