강의

N
챌린지

멘토링

N
클립

커뮤니티

BEST
AI Technology

/

Deep Learning & Machine Learning

Deep Learning and PyTorch Bootcamp for Beginners (Easy! From Basics to ChatGPT's Core Transformer) [Data Analysis/Science Part 3]

This is a newly designed course that allows you to gradually learn the mathematics, theory, PyTorch-based implementation, transfer learning, and GPT's core transformer needed to understand deep learning, based on the instructor's own failed experiences when first learning deep learning.

(4.9) 93 reviews

1,676 learners

Level Basic

Course period Unlimited

  • funcoding
Deep Learning(DL)
Deep Learning(DL)
PyTorch
PyTorch
Machine Learning(ML)
Machine Learning(ML)
Artificial Neural Network
Artificial Neural Network
Vision Transformer
Vision Transformer
Deep Learning(DL)
Deep Learning(DL)
PyTorch
PyTorch
Machine Learning(ML)
Machine Learning(ML)
Artificial Neural Network
Artificial Neural Network
Vision Transformer
Vision Transformer
News list
funcoding님의 프로필 이미지
funcoding

[Fun Coding] Deep Learning Course Update Notifications

Hello. This is Dave Lee from Fun Coding.

I need to translate this Korean text to English while preserving any markers and following the guidelines. The text is: "다름이 아니라 딥러닝 강의에서 수정 사항이 있어서 가볍게 공유를 드립니다." This is a polite, formal Korean sentence. Let me break it down: - "다름이 아니라" - a polite opening phrase meaning "it's nothing special but" or "I'm reaching out because" - "딥러닝 강의에서" - "in the deep learning course/lecture" - "수정 사항이 있어서" - "there are some corrections/modifications" - "가볍게 공유를 드립니다" - "I'm sharing (this) briefly/lightly" (polite form) Translation: I

There was an error in the test_loss calculation part during test evaluation in the "Applying CNN to MNIST Classification Problem" chapter. We apologize for the inconvenience. Therefore, we have updated the related materials with the following context.

nn.NLLLoss() has reduction='mean' applied by default, so it returns the average loss of samples within each mini-batch. Therefore, the test_loss accumulated through the for loop becomes "the sum of mini-batch average values".

In the existing code, this value was divided by the total number of samples (len(test_batches.dataset), e.g., 10000), which results in a much smaller value than the actual loss. For correct calculation, it should be divided by the number of mini-batches (len(test_batches), e.g., 79).

In other words, since it's the sum of mini-batch averages, dividing by the number of mini-batches accurately calculates the overall average loss.

I understand you want me to translate "수정된 파일:" from Korean to English. Modified files:

  • I understand you want me to translate the content of "12_CNN_MNIST.ipynb". However, I don't see any Korean text content provided to translate.

  • # Translation Analysis This appears to be a filename for a Jupyter notebook file. According to the translation guidelines: 1. Filenames should generally be preserved as they are technical elements 2. The components are

  • # Translation Analysis This appears to be a filename for a Jupyter notebook. According to the technical requirements, I should preserve filenames, code elements, and technical terms unchanged. # Translation 12_CNN_MNIST_GPU_DROPOUT.ipynb

I have completed the modifications to the Jupyter Notebook materials, so if you download the course materials again, you can study with the updated code.

Thank you.

Comment