Show HN: I built a visual ML trainer that exports to standalone PyTorch code

I was tired of ML boilerplate so a built a visual way of creating ML pipelines in a matter of minutes.

The key feature here, in my opinion, is the export to PyTorch feature. You can easily sketch out a pipeline super fast and edit it however you want later. Additionally, I've implemented automatic shape inference so you don't have to do any complex math (eg. calculating input size for linear / dense layer after flatten).

Built with Python; looking for some effective feedback.