VGG Model Training and Evaluation

This repository provides tools for training and evaluating VGG models. It includes scripts to train VGG models on a specific dataset and then evaluate/test the model.

Table of Contents

1. Setup
2. Usage

Setup

Note: This is only for GPU

1. Create Python 3.10 Environment

conda create -n [environment name] python=3.10 -y
conda activate [environment name]

2. Clone this repository.

https://github.com/hyokonbanwa/jikken3_vision_recognition
# If you want to clone VGG_Demo (for jetson) at the same time, you can use below instead of above.
# git clone --recurse-submodules https://github.com/ia-gu/jikken3_vision_recognition
cd jikken3_vision_recognition

3. Install required packages.

pip install -r requirements.txt

Usage

1. Place your data in the appropriate directory.
2. Train and test the model.

python main.py

3. (Optional) Only test the model.

cd src
python test.py

4. Do online demo

Note: This is only for jetson.

Access to jetson_camera_demo