metadata
{}
This model above was trained for some research. It was only trained on some 50k samples of text rows no longer than 1.2k length, and isn't good for any kind of deployment. It was trained from scratch with custom configuration. For 2 epochs.