feat: add focal loss

[stephantul]

This PR adds a focal loss to the classification trainer.

[greptile-apps]

Confidence Score: 4/5

[Medium risk] Adds focal loss implementations to the training classifier.

Fix the nonfinite gradients for fractional focal_gamma before merging.

_{Reviews (1) · Last reviewed commit: "feat: add focal loss" · Reviewed by Greptile}

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

Files with missing lines	Coverage Δ
model2vec/train/classifier.py	100.00% <100.00%> (ø)

🚀 New features to boost your workflow:

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