- Intro to Pytorch: The purpose of this notebook is to give you a general understanding of how to use the PyTorch Python package for writing, training and analysing neural networks.
- Linear Classifier: Train a multiclass linear classifier on the CIFAR10 dataset in Keras.
- Using Word2Vec Embeddings: News Category Classification using Word2Vec embeddings with two Python libraries, scikit-learn and gensim.
- Image Classification with CNN: Use convolutional neural networks, from development to training and testing. Plot feature maps and filters during the training process and testing. Explore methods of improving performance on a network. The details can be found in this article Image Classification with Convolutional Neural Networks - ImageNet10.
- Image Caption Generation: Do text pre-processing and text embeddings with an image to text model, compare the performance and usage of RNNs versus LSTMs as sequence generators. The details can be found in this article Image Caption Generation with BLEU
- Perceptron: Compare a simple perceptron and a simple network built by myself to a network model by Keras. The details can be found in this article Perceptron Algorithm and Backpropagation.
- CIFAR-10: The CIFAR-10 dataset consists of 60000 32x32 colour images in 10 classes, with 6000 images per class. There are 50000 training images and 10000 test images.
- ImageNet10: A subset of images from the ImageNet dataset, which contains 9000 images belonging to ten object classes.
- Flickr8k: This repository contains a copy of machine learning datasets used in tutorials on MachineLearningMastery.com.
- Iris: This is perhaps the best known database to be found in the pattern recognition literature. The data set contains 3 classes of 50 instances each, where each class refers to a type of iris plant.
- ResNet18: Use in Image_Classification_with_Convolutional_Neural_Networks_ImageNet10.ipynb
- ResNet152: Use in Image_Caption_Generation.ipynb
For most .ipynb
files, you can directly open it in Colab to run.
As for Image Caption Generation.ipynb, Flickr8k dataset is very large, so if you want to work on Colab, it is recommended to download the zip files, unzip them, and then upload all the files to your Google Drive. This initial upload may take a while, but from then on you will only need to mount your Drive every time you start a new Colab session and the data will be immediately accessible. Mounting only takes a few second. Do not forget to replace the path with your own root data directory of your Google Drive.
As for .py
files in the Perceptron folder, especially the perceptron.py
. You have to open terminal to type: python perceptron.py <class> with/without
, the class can be setosa
, versicolor
and verginica
. And with
means with learning rate 0.01, without
means without using learning rate.
This repository is licensed under Creative Commons Attribution 4.0 International License (CC BY 4.0).
本作品采用知识共享署名 4.0 国际许可协议进行许可。