Fix mapping for legacy topdown id models

[gitttt-1234]

This PR fixes the weight mapping for legacy top-down ID models. Previously, the legacy mapper didn’t account for linear layers and failed to transpose their weight shapes to match PyTorch’s convention. This update ensures that linear layer weights are correctly transposed and mapped from TensorFlow to PyTorch.

[codecov]

Codecov Report

❌ Patch coverage is 92.30769% with 1 line in your changes missing coverage. Please review.
✅ Project coverage is 93.44%. Comparing base (ff91433) to head (c884e5d).
⚠️ Report is 43 commits behind head on main.

Files with missing lines	Patch %	Lines
sleap_nn/legacy_models.py	92.30%	1 Missing ⚠️

❌ Your patch status has failed because the patch coverage (92.30%) is below the target coverage (95.00%). You can increase the patch coverage or adjust the target coverage.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #339      +/-   ##
==========================================
- Coverage   95.28%   93.44%   -1.84%     
==========================================
  Files          49       49              
  Lines        6765     7005     +240     
==========================================
+ Hits         6446     6546     +100     
- Misses        319      459     +140     

☔ 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.