Skip to content

Training neural networks for inverse design of nanophotonic gratings.

Notifications You must be signed in to change notification settings

FiodarM/InvDesignNet

Repository files navigation

Neural Network for Inverse Design in Nanophotonics

This project demonstates training a neural network for inverse design of nanophotonic gratings. The project is inspired by the work of Liu et al. in ACS Photonics journal.

Generating Data

  • To be able to run DNN training scripts, you should have dataset dataset.npz file in the project root folder.
  • If you want to use pre-generated dataset, you can download the dataset file from here. Note: The file size is ~1 GB.
  • In case you wish to generate the dataset from scratch, you can run either produce_data.py python script or produce_data.ipynb Jupyter notebook.

Forward Model

Inverse Model

About

Training neural networks for inverse design of nanophotonic gratings.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published