# CSE512 — Data

来源：https://www.kaggle.com/competitions/cse512hw3/data

本地 Meta Kaggle 快照；不是本次实时页面。

# Dataset Description

 <h2>File descriptions</h2>
<ul>
<li><strong>Train_Features.pkl</strong> - training set features</li>
<li><strong>Val_Features.pkl</strong> - Validation set features</li>
<li><strong>Test_Features.pkl</strong> - Test Set features</li>
<li><strong>Train_Labels.csv</strong> - training set labels</li>
<li><strong>Val_Labels.csv</strong> - Validation set labels</li>
<li><strong>Train_images96</strong> - Training set images</li>
<li><strong>Val_images96</strong> - Validation set images</li>
<li><strong>Val_Features.csv</strong> - Validation set features</li>
<li><strong>Train_Features.csv</strong> - Train set features </li>
<li><strong>Test_Features.csv</strong> - Test set features </li>
</ul>
<p>
To load the pickle files in Python 2.7, try:
<li>import pickle</li>
<li>TrainFileName = 'Train_Features.pkl' </li>
<li>Train_Features = pickle.load(open(TrainFileName,'rb'))</li> `
</p>
<p> In case you are using Matlab, you can read the csv Feature file as:
<li>name = 'Val_Features.csv'</li>
<li>fid = fopen(name,'r');</li>
<li>while ~feof(fid)</li>
    <li>tline = fgetl(fid);</li>
 <li>%use strplit to split the line, first cell contains the id, rest 512 cells contain the 512-d feature</li>
    <li>disp(tline)</li>
<li>end</li>
<li>fclose(fid);</li>
</p>