DrPCE-Net: Differential Residual PCE Network for Characteristic Prediction of Transistors

Introduction

🤔🤔🤔

This is the code implementation of the paper “DrPCE-Net: Differential Residual PCE Network for Characteristic Prediction of Transistors”.

The structure of the project is as follows：

-- DrPCE-Net (the main method proposed in the paper)

-- PCE (the implementation of Polynomial Chaoes Expansion)

-- baseline1 (the implementation of Y-model referred in the paper)

-- baseline2 (the implementation of W-model referred in the paper)

-- data (the data used in the experiments including three datasets: circle, rectangle, triangle)

-- net (the implementation of DNN referred in the paper)

-- plot (the source code of plotting)

-- results (the results of partial experiments)

-- tools (the implementation of some basic function)

Requirements

🥸🥸🥸

This project bases on Python 3.10.8.

More information about packages can be found in requirements.txt.

Usage

😇😇😇

Taking the DrPCE-Net as an example, you can train and test the model using the following command:

cd DrPCE-Net

python shell.py

By executing it , you will get a directory of predciting results, and you can view the report.txt to get an overview of the total prediction.

Docs

🥳🥳🥳

The paper is published on IEEE Transactions on Electron Devices, and you can find by clicking the following link:

DrPCE-Net: Differential Residual PCE Network for Characteristic Prediction of Transistors