Skip to content

tahsdj/yaliang_CNN

Repository files navigation

yanliang_CNN

Introduction

This is project for training a cnn model to distinguish between images of chess boards, green bean cakes and manuscript papers

Here is the link of code on colab: https://colab.research.google.com/drive/16lSezFOMhdKUwjCxgR1DQejs89GWZ2Rh

Motivation

"" 「啊,好像棋盤似的。」 「我看倒有點像稿紙。」我說。 「真像一塊塊綠豆糕。」一位外號叫「大食客」的同學緊接著說。

人總會去尋求自己喜歡的事物,每個人的看法或觀點不同,並沒有什麼關係,重要的是──人與人之間,應該有彼此容忍和尊重對方的看法與觀點的雅量。

──出自雅量 ""

機器學習也一樣,每個模型的觀點也不同,有的預測是綠豆糕有的預測像稿紙而有的分類成綠豆糕。因此當分類準確率低的時候不要氣餒,我們要培養雅量學習尊重各個模型的結果。

Dataset

  • 3 classes (chessboard, script paper, green bean cake)
cheesboard script paper green bean cake
  • Training data training: 606 samples
  • Test data: 202 samples
  • Image size: 200x200x3

Model

  • Network in Network

Results

  • Test acc: 98.02%
  • Prediction:

About

This is a CNN for distinguishing between images of chess boards, green bean cakes and manuscript papers

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published