Option 1: Lumi Colab¶
Run FDTD notebooks for free on Lumi Colab! Similar to Google Colab but preloaded with Lumi FDTD, Copilot AI and persistent storage. Has runtime and size restrictions and excludes inverse design. Upgrade to Lumi Colab Enterprise tier for unrestricted usage: info@luminescentai
Option 2: Local Python frontend + Lumi Colab backend¶
Run Python notebooks locally. lumi.solve uploads problem files to Lumi Colab server and downloads solution files. Same tiers apply
Prereqs¶
Python (Anaconda distro recommended for Windows users)
VS Code or another IDE
On a Linux VM you may need
sudo apt-get update
sudo apt install -y harminv libxi6 libxxf86vm1 libxrender1 libgl1 libglx-mesa0 libsm6 libice6 libxext6 libglib2.0-0 libglu1-mesa libxcursor1 libxft2 libxinerama1Frontend¶
pip install luminescent
Backend¶
Set environmental variables with YOUR_TOKEN from Lumi Colab. A copy of each run is kept in your Lumi Colab home at ~/local/runs/<run folder name>.
On Linux:
export LUMI_API=https://api.luminescentai.com
export LUMI_TOKEN=YOUR_TOKENOn Windows (PowerShell), then restart your terminal or IDE:
setx LUMI_API https://api.luminescentai.com
setx LUMI_TOKEN YOUR_TOKENOption 3: fully local¶
Executables can be compiled for enterprise customers above major spend threshold.