Click here for the previous article
Kaggle Kernel Summary [Table Time Series Data]
In this article, I’d like to know something like a standard for each data type in the Kaggle competition, including myself! I will write for people like. Also, I think it would be good if it could be a hint when accuracy does not come out regardless of the competition.
This time, we will look at various kernels, not limited to competitions.
When it comes to visualization in the case of images,
–Visualization of CNN layer by layer
–Visualization of image contribution
–Visualization (display) of the dataset image itself
Is it like that? (If there is anything else, please let me know in the comments)
Compared to other data types, it seems that the width is narrow, so this time we will focus on the method.
Introduction to CNN Keras – 0.997 (top 6%)
Ordinary CNN, model architecture seems to be exquisite
I am using LSTM-CNN
CNN Architectures : VGG, ResNet, Inception + TL
Transfer Learning is combined using VGG16, VGG19, InceptionNet, Resnet, XceptionNet, etc.
How to choose CNN Architecture MNIST
I’m trying different architectures
I would like to update it from time to time. Wish you a good year!
Read more here: Source link