Skip to content

It is a simple project where vehicle classification in malaysia for research purpose. This project will be use in deep learning apporach include efficientnet and inception

Notifications You must be signed in to change notification settings

wallacelkw/Vehicle-Classfication-using-Deep-Learning

Repository files navigation

Vehicle Classification

This project is use ChatGPT create a Inception and EfficientNet's code.

In this method we train a convolutional neural network with 6 classes which contain Bus, Cars, Motorcycles, Pickup, Truck and Van in Malaysia. The model it use Inception v1 and EfficientNet b3.

Source Code Folder

Main code

Setup environment pip install requirements.txt

To run the code for training, one can use the following command:
python train.py --modelname inception

To test the model one can use the following command: python test.py --modelname inception

The Model Architecture of Inception and Efficientnet will be in efficient.py and inception.py

All the information of function will be in utils.py

EarlyStopping class will be stored in pytorchtools.py

Preprocessing File-sub folder

All preprocessing used jupyter lab to perform:

all the processing will be store in these file which include image augmentation, convert to h5 file, and train test split.

  1. Augmentation image_augmentation.ipynb
  2. Web Scraping scrap_from_website.ipynb
  3. Train, validation, test splitting train_test_split.ipynb
  4. Convert into h5 file for machine learning convert2h5.ipynb

Result sub folder

The result contain Inception and Efficient classification report, training and validation 's loss and accuracy, confusion matrix and the checkpoint file of the trained model.

Screenshot

EfficientNet

Loss and Accuracy

Classification Report of Inception

InceptionNet

Loss and Accuracy

Classification Report of Inception

About

It is a simple project where vehicle classification in malaysia for research purpose. This project will be use in deep learning apporach include efficientnet and inception

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published