Deep Learning with Python for Image Classification

在本课程中，您将学习使用 Python 和 PyTorch 进行深度学习以使用预训练模型进行图像分类。图像分类是一项计算机视觉任务，用于识别输入图像并使用机器学习技术预测图像的单标签或多标签作为输出。

您将使用 Google Colab 笔记本编写使用深度学习模型进行图像分类的 Python 代码。

您将了解如何将 Google Colab 与 Google Drive 连接以及如何访问数据。

您将使用不同的变换执行数据预处理，例如图像调整大小和中心裁剪等。

您将使用 Python 的深度学习模型执行两种类型的图像分类、单标签分类和多标签分类。

您将能够学习迁移学习技术

1. 通过 FineTuning 模型进行迁移学习。

2. 通过使用模型作为固定特征提取器来迁移学习。

您将学习如何执行数据扩充。

您将学习如何加载数据集和数据加载器。

您将学习微调深度 Resnet 模型。

您将学习如何使用 Deep Resnet 模型作为固定特征提取器。

您将学习超参数优化和结果可视化。

在单标签分类中，当您将输入图像提供给网络时，它会预测单个标签。在多标签分类中，当您将输入图像提供给网络时，它会预测多个标签。您将学习深度学习架构，例如 ResNet 和 AlexNet。ResNet 是一种用于图像分类和识别的深度卷积神经网络。专为分类任务设计的 ResNet 网络架构，在包含 1000 个类别的自然场景 imageNet 数据集上进行训练。深度残差网络在 ILSVRC 2015 分类挑战赛中获得第一名。Alexnet 是一个在 ImageNet 数据集上训练的深度卷积神经网络，用于将图像分类为 1000 类。它有五个卷积层，后面跟着最大池化层和 3 个全连接层。AlexNet 赢得了 ILSVRC 2012 分类挑战。您将使用 ResNet 和 AlexNet 深度学习模型执行图像分类。深度学习社区从这些开源模型中受益匪浅，其中预训练模型是计算机视觉和深度学习研究快速发展的主要原因。

Learn Deep Learning & Computer Vision for Image Classification using Pre-trained Models with Python using Google Colab

Instructors: Computer Science & AI School
19 sections • 21 lectures • 1h 31m
Video: MP4 1280×720 44 KHz | English + Sub
Updated 11/2022 | Size: 795 MB

What you’ll learn

Learn Image Classification using Deep Learning PreTrained Models
Learn Single-Label Image Classification and Multi-Label Image Classification
Learn Deep Learning Architectures Such as ResNet and AlexNet
Write Python Code in Google Colab
Connect Colab with Google Drive and Access Data
Perform Data Preprocessing using Transformations
Perform Single-Label Image Classification with ResNet and AlexNet
Perform Multi-Label Image Classification with ResNet and AlexNet
Learn Transfer Learning
Dataset, Data Augmentation, Dataloaders, and Training Function
Deep ResNet Model FineTuning
ResNet Model HyperParameteres Optimization
Deep ResNet as Fixed Feature Extractor
Models Optimization, Training and Results Visualization

Requirements

Deep Learning with Python and Pytorch is taught in this course
A Google Gmail account to get started with Google Colab to write Python Code

Description

In this course, you will learn Deep Learning with Python and PyTorch for Image Classification using Pre-trained Models. Image Classification is a computer vision task to recognize an input image and predict a single-label or multi-label for the image as output using Machine Learning techniques.

You will use Google Colab notebooks for writing the python code for image classification using Deep Learning models.

You will learn how to connect Google Colab with Google Drive and how to access data.

You will perform data preprocessing using different transformations such as image resize and center crop etc.

You will perform two types of Image Classification, single-label Classification, and multi-label Classification using deep learning models with Python.

You will be able to learn Transfer Learning techniques

1. Transfer Learning by FineTuning the model.

2. Transfer Learning by using the Model as Fixed Feature Extractor.

You will learn how to perform Data Augmentation.

You will learn how to load Dataset, Dataloaders.

You will Learn to FineTune the Deep Resnet Model.

You will learn how to use the Deep Resnet Model as Fixed Feature Extractor.

You will Learn HyperParameters Optimization and results visualization.

In single-label Classification, when you feed input image to the network it predicts single label. In multi-label Classification, when you feed input image to the network it predicts multiple labels. You will Learn Deep Learning architectures such as ResNet and AlexNet. The ResNet is a deep convolution neural network proposed for image classification and recognition. ResNet network architecture designed for classification task, trained on the imageNet dataset of natural scenes that consists of 1000 classes. Deep residual nets won the 1st place on the ILSVRC 2015 Classification challenge. Alexnet is a deep convolution neural network trained on ImageNet dataset to classify the images into 1000 classes. It has five convolution layers followed by max-pooling layers, and 3 fully connected layers. AlexNet won the ILSVRC 2012 Classification challenge. You will perform image classification using ResNet and AlexNet deep learning models. The Deep Learning community has greatly benefitted from these open-source models where pre-trained models are a major reason for rapid advancements in the Computer Vision and deep learning research.

Who this course is for
Deep Learning enthusiasts interested to learn with Python and Pytorch
Students and researchers interested in Deep Learning for Image Classification