Table of Contents

• Introduction
• Experiment Result
• Installation
• Usage
• Models
• Project Structure
• RelativeProject
• Contributing
• License

Introduction

This project provides a framework for training and evaluating CNN models on the CIFAR-10 and CIFAR-100 datasets. It includes various models such as ResNet, VGG, and MobileNetV2, and utilizes PyTorch's deep learning library for efficient training and evaluation.

Experiment Result

Dataset	Network	Params (M)	Top1 Acc (%)	Epoch
Cifar100	MobilenetV2	2.412	79.32	200
Cifar100	VGG16	34.015	76.41	200
Cifar100	Resnet101	42.697	83.41	200
Cifar10	MobilenetV2	2.297	95.82	200
Cifar10	VGG16	33.647	95.26	200
Cifar10	Resnet101	42.513	-	-

Installation

To install the required dependencies, follow these steps:

1. Clone this repository:

git clone https://github.com/MIC-Laboratory/Pytorch-Cifar.git

2. Navigate to the project directory:

cd Pytorch-Cifar

3. Install the dependencies:

pip install -r requirements.txt

Usage

To use this project, follow these steps:

1. Configure the training parameters:

   • Open the config.yaml file and set the desired hyperparameters such as batch size, learning rate, and number of epochs.
   • Modify other settings such as model type, dataset path, and output directory if needed.

2. Train the model:

   • Run the training script:

     python Training.py

   • Monitor the training progress and observe the log output.

Models

This project includes the following pre-defined models:

• Resnet101
• VGG16
• MobilenetV2

These models are defined in separate Python files in the Models directory.

Pretrain Weight

Pretrained Weight for Cifar10 and Cifar100 download here

Project Structure

The structure of the project is as follows:

Pytorch-Cifar
├── config.yaml
├── LICENSE
├── Models
│   ├── Mobilenetv2.py
│   ├── Resnet.py
│   └── Vgg.py
├── README.md
├── requirements.txt
├── Training.py
│── Weapon
│   └── WarmUpLR.py

• config.yaml: Configuration file containing hyperparameters and settings for training the models.
• Training.py: Script for training and testing the CNN models using PyTorch.
• Models directory: Contains implementation files for different CNN models.
  • Resnet.py: Contains code for the ResNet model.
  • Vgg.py: Contains code for the VGG model.
  • Mobilenetv2.py: Contains code for the MobileNetV2 model.
• Weapon directory: Contains code related to learning rate scheduling.
  • WarmUpLR.py: Defines the WarmUpLR class for implementing learning rate warm-up. README.md: Documentation file providing an overview of the project and instructions for usage. This structure tree provides an overview of the organization of the project files.

Relative Project

Pruning-Engine

On-device-CNN

Contributing

We welcome contributions from the community to improve this project. If you encounter any issues or have suggestions for enhancements, please feel free to submit a pull request or open an issue on the GitHub repository.

License

This project is licensed under the MIT License.