This project is a sandbox project! This projects will be used for experiments!

Image Classifier using TensorFlow

How to Use

Installation

1. This project can be run only linux environment.
2. This project is recommended to run on virtual environment include docker.
3. Command pip install -r requirements.txt to install dependency.

Commands

• Train: python3 train.py option=<configuration file (optional)>

  • Used to obtain weights.
  • The option argument for indivisual training options.

• Predict: python3 classify.py weights=path/to/checkpoint_path/checkpoint_name image=path/to/image(s/dir) epoch=[best|last|<an epoch number>](optional) output_format=[csv|stdout](optional)

• Plot graph: python3 plot.py path=path/to/checkpoint_path/checkpoint_name

  • Used to plot graph of train result.
  • The graphs will be saved onto path/to/checkpoint_path/checkpoint_name with plot.png.

• Export: python3 export.py path=path/to/checkpoint_path/checkpoint_name epoch=[best|last|<an epoch number>] image_size=<image_size>

  • Export weights to saved_model and TFLite
  • The weights files will be saved onto path/to/checkpoint_path/checkpoint_name/export.

• Select specific GPU (example): CUDA_VISIBLE_DEVICES=[cpu|<GPU numbers (example: 0,1,2,3)>] python3 train.py option=example.yaml

  • The default uses all GPUs.
  • If you have multiple GPUs and want to select specific GPU, you have to use this command form or set the CUDA_VISIBLE_DEVICES variable.

Options

Option application order

1. cfg/settings.yaml (default settings)
2. option.yaml (user global settings)
3. configuration file (copied from option.yaml)

option.yaml

• Dataloader options
  • Image crop options
  • Image translate options
  • Color adjustment options
  • Quality adjustment options
  • Other dataloader options
• Train options
  • Checkpoint path
  • Training options (optimizer, loss, etc.)
  • Learning rate scheduler options
  • Model option

Model Configurations

• ResNet Models

  • resnet/18.yaml
  • resnet/54.yaml
  • resnet/102.yaml
  • resnet/150.yaml
  • resnet2l/15.yaml
  • resnet2l/54.yaml
  • resnet2l/110.yaml

• CSPResNet Models

  • cspresnet/18.yaml
  • cspresnet/54.yaml
  • cspresnet/102.yaml
  • cspresnet/150.yaml
  • cspresnet2c/18.yaml
  • cspresnet2c/54.yaml
  • cspresnet2c/102.yaml
  • cspresnet2c/150.yaml
  • cspresnet2l2c/15.yaml
  • cspresnet2l2c/54.yaml
  • cspresnet2l2c/110.yaml
  • cspresnet2l3c/15.yaml
  • cspresnet2l3c/54.yaml
  • cspresnet2l3c/110.yaml

• Other Models

  • vgg16.yaml
  • googlenet.yaml

• Details

  • The ResNet models are implemented of 3-layers ResNet blocks with bottleneck.
  • The ResNet2L models are implemented of 2-layers ResNet blocks without bottleneck.
  • The CSPResNet and CSPResNet2L3C models are implemented of 3-Conv layers to implementing CSPNet structure.
  • The CSPResNet2C and CSPResNet2L2C models are implemented of 2-Conv layers to implementing CSPNet structure.

Experiment

Experimental Environment

• CPU: AMD Ryzen 7900 @ 4.0GHz; 90W
• Memory: DDR5 64GB
• GPU: RTX 4060 Ti 16GB @ 2.5GHz; 130W x2
• OS: Ubuntu 24.04 (docker image: nvidia/cuda:12.2.2-cudnn8-devel-ubuntu22.04)
• Python: CPython 3.12
• TensorFlow version: 2.20.0