![[개정판] 파이썬 머신러닝 완벽 가이드강의 썸네일](https://cdn.inflearn.com/public/courses/324238/cover/7e380aa0-48ba-4ee7-a6b2-8da7900568d6/324238-eng.png?w=420)
[개정판] 파이썬 머신러닝 완벽 가이드
권 철민
이론 위주의 머신러닝 강좌에서 탈피하여 머신러닝의 핵심 개념을 쉽게 이해함과 동시에 실전 머신러닝 애플리케이션 구현 능력을 갖출 수 있도록 만들어 드립니다.
초급
Python, 머신러닝, 통계
From core theories of deep learning and CNN to various CNN model implementation methods, and practical deep learning development know-how through real-world problems, If you want to become a deep learning CNN technology expert based on Pytorch, join us in this lecture :)
528 learners
Understanding the Core Technical Elements and Major Models of Deep Learning and CNNs
Understanding Model Construction and Training Process Using Pytorch
Image classification using CNN and performance optimization techniques
Improving Performance Using Pretrained Models and Fine-Tuning
Various image augmentation techniques and methods to improve model performance using them.
Practical deep learning development workflow, including image preprocessing, data processing, model creation, optimal performance improvement, and performance evaluation.
✅
Already proven course. Renewed with Pytorch .
Deep Learning CNN Complete Guide (TFKeras) with 2,000+ students and a rating of 5.0
We're back with a Pytorch version with even more enhanced practice and the latest trends.
✅
40 hours, 320 pages of deep understanding
Learn the basic principles of CNN, major models, and performance optimization techniques systematically. Through this, you can gain a deep understanding of CNN and acquire the ability to optimize models on your own.
✅
Experience the deep learning practical workflow
From image preprocessing to model creation, performance improvement, fine-tuning, and evaluation! You can gain practical experience by following the deep learning development process.
✅
A lecture that delves deep into Pytorch
You will gain a clear understanding of the key concepts of PyTorch and be able to freely utilize it from model building to optimization through various practical exercises.
Computer vision is rapidly spreading in cutting-edge technologies such as autonomous driving, smart cities, medical imaging diagnosis, and AR/VR.
Deep learning-based computer vision technology is accelerating industrial automation and technological innovation , and will continue to create demand and career opportunities in the future.
Tesla is building a camera-based autonomous driving system, revolutionizing the existing sensor-centric approach. BMW is introducing AI-based vision systems to smart factories to automate manufacturing processes. In addition, Amazon is using logistics robots with computer vision technology to improve logistics processing speed by more than 40%.
The reason why CNN (Convolutional Neural Network) is fundamental and core in computer vision is because it provides a structure optimized for image processing and powerful performance that serves as the basis for actual commercialized models compared to other models .
✅ The model that best finds the features of the image
✅ Core base of many commercial models such as YOLO, Mask R-CNN, ResNet, etc.
✅ High performance with little data through transfer learning, can be quickly applied to practice
As computer vision technology is rapidly developing these days, it is essential to accurately understand and utilize CNN models that can be used immediately in practice. In this course, you will learn step-by-step from the principles of CNN to practice based on PyTorch, and build practical skills that can be applied immediately in practical projects.
We will install the core fundamental knowledge of deep learning and CNN in your head through in-depth theory and practice.
We will help you implement CNN applications by freely utilizing Pytorch through detailed explanations and practical training on the core frameworks that make up Pytorch.
You can freely implement CNN image classification models through various datasets and practical problems, and learn the latest performance tuning techniques such as Augmentation, Learning Rate optimization, and EfficientNet utilization.
In order to use CNN for applications that extend beyond image classification models, it is important to understand the development process and core technologies of modern CNN models. To this end, we will explain in detail the architecture and characteristics of major CNN models such as AlexNet, VGGNet, GoogLeNet (Inception), and ResNet, as well as their implementation methods at the source code level .
From learning gradient descent through direct implementation, to error backpropagation, activation functions, loss functions, and optimizers, it is composed of various materials, practice codes, and detailed explanations so that you can acquire a solid foundation of deep learning knowledge.
From handling Tensors, nn.Modules, submodules, Layers, modularization, and Train Loops, you can easily understand the core elements for configuring Pytorch's network models step by step through detailed explanations and hands-on practice.
We have designed the lecture to help you easily understand the main components of CNN with various visual materials and practical exercises, and to help you naturally understand more advanced performance improvement techniques.
Beyond the basic usage of constructing a Dataset and calling a DataLoader, we will detail the main parameters and interaction mechanisms of the Dataset and DataLoader.
You can configure a custom model and fine tuning by utilizing a pretrained model based on torchvision and timm.
We will learn how to implement Model Checkpoint and Early Stopping functions and apply various evaluation indicators using Torchmetrics.
We will teach you how to use libraries such as torchvision transform and Albumentations, as well as various Augmentation techniques and CutMix applications through hands-on practice. We will also talk about how to improve model performance by applying Augmentation.
We will explain in detail the architecture and characteristics of important core CNN models such as AlexNet, VGGNet, GoogLeNet (Inception), and ResNet, as well as the implementation of these models at the source code level. We will also help you understand the models with easy theoretical explanations for EfficientNet V1 and V2.
Through comprehensive practice problems, you can acquire differentiated practical skills at the expert level by applying the latest CNN models and various model optimization methods covered so far.
The practice environment is performed using the notebook kernel provided by Kaggle. After signing up for Kaggle, if you select the Code menu, you can use the P100 GPU for 30 hours a week for free based on the Jupyter Notebook environment similar to Colab.
Who is this course right for?
For those who want to greatly improve their skills in deep learning and CNN.
For those who want to confidently improve their PyTorch skills
For those who want to utilize deep learning image classification models in the field of computer vision
For those preparing for image classification competitions on Kaggle or Dacon
Need to know before starting?
Basic understanding of the first half of "Perfect Guide to Python Machine Learning," from the beginning up to Section 4, Evaluation.
26,928
Learners
1,367
Reviews
4,010
Answers
4.9
Rating
14
Courses
(전) 엔코아 컨설팅
(전) 한국 오라클
AI 프리랜서 컨설턴트
파이썬 머신러닝 완벽 가이드 저자
All
217 lectures ∙ (41hr 33min)
Course Materials:
6. Section Overview
00:30
18. Using argmax
07:35
20. Various indexing
11:56
22. Section Overview
02:22
42. Section Overview
01:31
55. Creating DataLoader
11:17
62. Optimizer Overview
07:24
84. Stride and Padding
11:50
88. Pooling
12:05
114. Weight Decay
07:45
127. Overview of timm
05:46
130. timm, a closer look
09:44
217. In conclusion,
01:09
All
30 reviews
5.0
30 reviews
Reviews 2
∙
Average Rating 5.0
Edited
5
현재 섹션9까지 들었는데 최고의 강의인 것 같습니다. 혹시 object detection 같은 advanced 강의는 언제쯤 올라올 예정인가요?
오, 칭찬 넘 감사드립니다. 후속 강의는 올해 가을 되기 전에 완료할 생각입니다.
Reviews 1
∙
Average Rating 5.0
Reviews 2
∙
Average Rating 5.0
Reviews 1
∙
Average Rating 5.0
Limited time deal ends in 00:58:01
$82,500.00
25%
$84.70
Check out other courses by the instructor!
Explore other courses in the same field!