fix: stop the trainer's tokenizer from padding

[stephantul]

The trainable model's copy of the tokenizer kept its padding setting, so tokenizers with padding enabled padded every text to the longest text in its tokenization chunk. Share the padding-disabling helper from StaticModel, which keeps the pad token but never pads.

[greptile-apps]

Confidence Score: 5/5

[Medium risk] Refactors tokenizer padding behavior in model initialization.

The PR appears safe to merge; no actionable issue was identified.

_{Reviews (1) · Last reviewed commit: "fix(train): stop the trainer's tokenizer..."}

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

Files with missing lines	Coverage Δ
model2vec/model.py	97.56% <100.00%> (+0.02%)	⬆️
model2vec/train/base.py	99.54% <100.00%> (+<0.01%)	⬆️

🚀 New features to boost your workflow:

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