Classification_Task

Multi-Neural Networks

The project was implemented aims to realize three types rock thin section images classification. And the method was optimized based on GoogLeNet、VGG16、MobileNetV2、ShuffleNetV2.

Contents

• 1.Environment

• 2.Preparation

• 3.Training

• 4.Demo

Environment

- Driver: NVIDIA-Linux-x86_64-470.63.01.run
- CUDA: cuda_11.4.2_470.57.02_linux.run
- Torch 1.7.0+
- python 3.8

Preparation

1. Prepare the dataset, we spyder the data from ScienceDB.

2. Prepare the training dataset, the data structure as follow.

    $Data/                           # RootPath  
    
    $Data/sub-class folder                 # include images.   
   

Training

1. Train your model on PASCAL VOC Format.

    cd $Classification_Task
    python3 googlenet-p-r-f1-lambda.py/googlenet-p-r-f1-lambda-rms.py/mobilenetv2-p-r-f1-lambda-rms.py
   

2. Train results, it will create '.pth' model, loss log file and evaluation log file.

    # It will create model definition files and save snapshot models in:
    #   - $Classification_Task/weights/'{}_{}.pth'
    # the loss log and evaluation log saved in:
    #   - $Classification_Task/log/'{}'/loss.txt'.format(lstrftime('%b%d-%H'))
    #   - $RDNet/log/'{}'/result.txt'.format(lstrftime('%b%d-%H')