InverTwin

Inverse design and optimization tutorials for RF systems

📦

Package Download

Complete package with data, notebooks, and .whl files

The package, which includes data, notebooks, and .whl files is available for download at:

⚠️ Note: This is an Older Version (2024)

Download link:

https://github.com/RFDigitalTwin/RFDigitalTwin.github.io/releases/download/InverTwinDemo/demo.zip
Download demo.zip

Installation

The supplementary code and package is tested on Linux 22.04 with GPU NVIDIA RTX A6000. The package, which includes data, notebooks, and .whl files is available for download at: demo.zip.

Create and activate conda environment:

conda create --name RFDT_demo -y python=3.8
conda activate RFDT_demo

Install requirements:

conda install -c conda-forge gcc=12.1.0
python -m pip install --upgrade pip
pip install torch==2.1.2+cu118 torchvision==0.16.2+cu118 --extra-index-url https://download.pytorch.org/whl/cu118
pip install numpy matplotlib opencv-python ipykernel
conda install -c "nvidia/label/cuda-11.8.0" cuda-toolkit

Download and install InverTwin package:

Download invertwin-0.0.1-cp38-cp38-linux_x86_64 from the demo package, then:

pip install invertwin-0.0.1-cp38-cp38-linux_x86_64.whl
📈

Local Non-convexity

Analyze optimization landscapes and escape strategies for local minima

🌐

Spatial Spectrum

Explore spatial spectrum analysis and frequency domain optimization

∇

Gradients

Understand gradient-based optimization and differentiable simulation

Getting Started

  1. 1Download the demo.zip package using the link above
  2. 2Extract the package and install the .whl files
  3. 3Run the interactive notebooks to explore inverse design techniques