# Copy of Data Challenge - Kernel methods (2023-2... — Data

来源：https://www.kaggle.com/competitions/data-challenge-kernel-methods-2023-2024-extension/data

本地 Meta Kaggle 快照；不是本次实时页面。

# Dataset Description

This page shows up underneath the data files. 

##File descriptions
- Xtr.csv - the training images.
- Xte.csv - the test images
- Ytr.csv - the image labels of the training images, in the same format as a submission file

Xtr.csv contains a matrix of size 5000 x 3072.  One row represents a color image of size 32 x 32 pixels. The first 1024 values represent pixel intensities on the red channel, then the next 1024 represent the green channel, and the last 1024 entries, the blue channel. To make sure that you are reading the data correctly, you need to find a way to visualize images. Note that images have been pre-processed, and do not appear as a natural image.

Similarly, Xte.csv contains 2000 test images. These are the ones you need to classify. The first row corresponds to image Id=1, then the last row corresponds to image Id=2000.

Ytr.csv contains a vector of labels corresponding to the training data, in the same format as a submission file. 
You can find more details in this format in the Overview/Evaluation section.