I made an app that utilizes Burn's LSTM for data forecasting

[Thor-x86]

First of all, many thanks for all contributors here. I'm happy to announce that I released a simple GUI app to predict upcoming data according to past historical data.

It uses a Long-Short Term Memory (LSTM) layer with sequential 1D input. Inspired from modern-lstm example. For 1 month of time-series data (daily interval), it will take around 15 minutes from training to predict on AMD Ryzen 7 3700U. Might be vary depending on hardware specs.

You can try yourself here: https://github.com/Thor-x86/project-tarot#readme

It's just a hobby and won't be big anyway. Despite of that, I welcome all feedback and suggestions.