EfficientAD

This repository is customized based on EfficientAD.

Example of Inference Visualization (juice bottle)

Loss history (juice bottle)

Key Customizations

1. Modularization:

   • The codebase has been modularized for easier management and scalability.

2. Training Process Update:

   • The training loop has been adjusted from an iteration-based approach to an epoch-based approach.

3. Loss Function Adjustment:

   • The penalty loss has been removed when calculating ( L_{st} ).

4. Visualization during Inference:

   • Added support for visualization during the inference phase.

Getting Started

Please refer to the original EfficientAD documentation for setup instructions. This repository retains compatibility with the original setup process.

Prerequisites

• Python 3.11
• Required libraries can be installed using:

  pip install -r requirements.txt

Usage

To train the model:

python train.py -s "{dataset directory}" -m "{model size}"

To run inference with visualization:

python inference.py -t "{testset directory}" -m "{model directory}