Programmer Group

Hello, everyone. Today, I'd like to share with you the derivation process of forward propagation in tensorflow 2.0 deep learning, using the mnist data set provided by the system. 1. Data acquisition First, we import the required library files and datasets. The imported x and y data are array types and need to be converted to tensor type tf.co ...

1, Multilayer perceptron model based on BP algorithm The multilayer perceptron using BP algorithm is the most widely used neural network so far. In the application of multilayer perceptron, the single hidden layer network shown in Figure 3-15 is the most common. In general, it is customary to call a single hidden layer feedforward network a th ...

preface: Two months ago, I wrote the complete steps of TextCNN (less than 60 lines of code), but did not take into account the subsequent engineering deployment and large amount of data (unable to load all into memory), so today I made a transformation and optimization according to the actual case. The operation steps of TextCNN can generally b ...

catalogue Plans completed this week Thesis Reading 1   Abstract (Abstract) 1. Introduction 2. Pseudolabel method for deep neural networks 2.1. Deep Neural Networks 2.2. Denoising auto encoder  2.3. Dropout   2.4. Pseudo label 3. Why could Pseudo-Label work? (why pseudo tags work) 3.1. Low density separation between classes ...

Summary of Li Hongyi's machine learning notes Course links Review The more complex the model, the lower the error on the testing data. error comes from two places: biasvariance f_star is f_ An estimate of hat, f_star may not be equal to f_hat, the distance may come from bias or variance Estimate the mean value of variab ...

  Video link: The final collection of PyTorch deep learning practice_ Beep beep beep_ bilibili This secondary implementation is a more complex neural network. Although the model looks very complex, in order to reduce code redundancy and improve code reusability, we can define the neural network with the same structure as a class to improv ...

Note: This is a machine learning practical project (with data + code). If you need data + complete code, you can get it directly at the end of the article. 1. Project background         For the fault diagnosis of semiconductor etcher, it is necessary to collect and obtain the data of the etching process o ...

Title Description Given the training data set (pictures of cats), let's build a simple neural network to identify cats. Dataset description There are 209 pictures in the training set, and the shape of each picture is (64, 64, 3) There are 50 pictures in the test set, and the shape of each picture is (64, 64, 3) classes stores two string data ...

preface shufflenetV1 is another direction of the development of convolutional neural network to lightweight, which is the lightweight network following Mobilenet. 1, Paper reading summary Paper address Tricks: application of group revolution on 1 * 1 convolution; channel shuffle improves information transmission between channels ...

Weight attenuation In the previous section, we observed the over fitting phenomenon, that is, the training error of the model is much smaller than its error on the test set. Although increasing the training data set may reduce over fitting, it is often expensive to obtain additional training data. This section describes a common method to ...