Fix metrics saving format to match SLEAP 1.4 and eliminate code duplication

[gitttt-1234]

Summary

• Standardized metrics saving to use SLEAP 1.4 format (np.savez_compressed with single "metrics" key)
• Eliminated code duplication by having train.py call run_evaluation() instead of duplicating evaluation logic
• Updated load_metrics() function to properly load the standardized format
• Cleaned up unused imports in train.py

Changes

• evaluation.py: Updated run_evaluation() to save metrics using SLEAP 1.4 format
• train.py: Refactored to use run_evaluation() and removed duplicated evaluation code
• test_evaluation.py: Updated tests to load metrics in the new format

Test plan

• All existing tests pass (252 tests: 248 passed, 4 xfailed, 4 skipped)
• Linting checks pass (black and ruff)
• CI/CD pipeline passes

🤖 Generated with Claude Code

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 93.84%. Comparing base (ff91433) to head (41022d8).
⚠️ Report is 62 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #371      +/-   ##
==========================================
- Coverage   95.28%   93.84%   -1.45%     
==========================================
  Files          49       49              
  Lines        6765     7179     +414     
==========================================
+ Hits         6446     6737     +291     
- Misses        319      442     +123     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

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