feat: remove lightning

[stephantul]

This PR removes all references to lightning and pytorch lightning. Scores are (on average) identical for classification, but the new training loop is faster. Since the core training loop is shared by all training variants, I have high confidence that this also works for multi-label and regression tasks.

[codecov]

Codecov Report

❌ Patch coverage is 99.31034% with 1 line in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
model2vec/train/trainer.py	99.01%	1 Missing ⚠️

Files with missing lines	Coverage Δ
model2vec/train/__init__.py	100.00% <ø> (ø)
model2vec/train/base.py	97.91% <100.00%> (-0.17%)	⬇️
model2vec/train/classifier.py	98.27% <100.00%> (+0.14%)	⬆️
model2vec/train/regression.py	100.00% <100.00%> (ø)
model2vec/train/similarity.py	100.00% <100.00%> (ø)
model2vec/train/utils.py	100.00% <100.00%> (ø)
model2vec/train/trainer.py	99.01% <99.01%> (ø)

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[Pringled]

Very nice! Two small comments but looks good to me