e4040-2022Fall-project-xswl

Basic Information:

• Project name (paper name): MobileNet: Efficient Convolutional Neural Networks for Mobile Vision Applications
• Group Menber: Yingqi Ma ym2926, Zhan Shu zs2584, Zihao Xiao zx2407

Location of Dataset:

• The dataset We used is cifar100 from keras.
• Website link: https://keras.io/api/datasets/cifar100/.

How to run the code:

• Since the data we use can be downloaded directly from the web, we wrote the code in jupyter notebook to load that dataset.
• You can directly download two files called E4040_project_fianlmodel.ipynb and mobilenet.py. Put them in a folder and upload it to your environment. Then you just need to open the E4040_project_fianlmodel.ipynb and run the code inside it.
• Opening the file called E4040_project_fianlmodel.ipynb

The name of the main jupyter notebook:

• E4040_project_fianlmodel.ipynb

Key functions of each file in the project:

• E4040_project_fianlmodel.ipynb: Our main jupyter notebook, contains all the processes of our project. It concludes Data loading, Data pre-processing, MobileNet building, Model training, Parameter tuning and Image drawing.
• mobilenet.py: This file contains the code of MobileNet architecture.
• E4040.2022Fall.XSWL.report.ym2926.zs2584.zx2407.pdf: Our subject report.
• Failed_Experiments.file - E4040_project-c100.ipynb: The jupyter notebook in this file contains the code of the model we originally built, but we abandoned it because the overfitting was so severe that it led to very poor results.

Organization of directories:

--README.md

--E4040_project_fianlmodel.ipynb

--mobilenet.py

--E4040.2022Fall.XSWL.report.ym2926.zs2584.zx2407.pdf

--Failed_Experiments.file-- E4040_project-c100.ipynb