Various Improvements around Flow Matching

[LarsKue]

While debugging Flow Matching, I fixed several smaller things, most notably:

• Flow Matching now uses OT by default
• OT skips most of its internal algorithm when tracing
• Flow Matching now uses Fixed Euler with 100 steps as the default
• keras_kwargs no longer relies on kwargs ending in _kwargs
• Sampling from a ContinuousApproximator is now significantly more robust against misuse
  • I am currently relying on the user passing num_datasets again, but I aim to remove this again before marking this as ready for review
  • EDIT: reverted this change

[codecov-commenter]

⚠️ Please install the to ensure uploads and comments are reliably processed by Codecov.

Codecov Report

Attention: Patch coverage is 65.95745% with 32 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
bayesflow/utils/tensor_utils.py	56.00%	22 Missing ⚠️
bayesflow/utils/optimal_transport/sinkhorn.py	20.00%	4 Missing ⚠️
.../experimental/continuous_time_consistency_model.py	0.00%	3 Missing ⚠️
bayesflow/approximators/continuous_approximator.py	77.77%	2 Missing ⚠️
bayesflow/networks/flow_matching/flow_matching.py	95.00%	1 Missing ⚠️

❗ Your organization needs to install the Codecov GitHub app to enable full functionality.

Files with missing lines	Coverage Δ
...flow/experimental/free_form_flow/free_form_flow.py	74.54% <100.00%> (ø)
bayesflow/utils/__init__.py	100.00% <ø> (ø)
bayesflow/utils/dict_utils.py	62.12% <100.00%> (+1.49%)	⬆️
bayesflow/utils/integrate.py	50.41% <100.00%> (-22.09%)	⬇️
bayesflow/workflows/basic_workflow.py	22.41% <ø> (-0.40%)	⬇️
bayesflow/networks/flow_matching/flow_matching.py	82.40% <95.00%> (-0.78%)	⬇️
bayesflow/approximators/continuous_approximator.py	69.16% <77.77%> (+17.76%)	⬆️
.../experimental/continuous_time_consistency_model.py	19.81% <0.00%> (ø)
bayesflow/utils/optimal_transport/sinkhorn.py	16.43% <20.00%> (+0.72%)	⬆️
bayesflow/utils/tensor_utils.py	53.37% <56.00%> (-7.11%)	⬇️

... and 2 files with indirect coverage changes

🚀 New features to boost your workflow:

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