Refactor: remove numpy and ml_dtypes dependencies

[ChaoWao]

Summary

Project now uses torch exclusively for tensor operations. numpy was only used as a fallback conversion path and isinstance check; ml_dtypes had no remaining call sites.

• Remove numpy fallback in _to_torch() and isinstance checks in code_runner.py and ci.py
• Ensure _to_torch() always returns cpu+contiguous tensors; remove redundant .cpu().contiguous() at call sites
• Remove pip install numpy/ml_dtypes from all CI workflow jobs; sim jobs now use pip install '.[test]'
• Add Dependencies section to python-packaging.md documenting the simpler vs simpler_setup dependency split
• Update getting-started.md install commands and docstrings

Testing

• Simulation tests pass (a2a3sim: 25 passed, 80 skipped, 0 failed)
• Hardware tests pass (if applicable)

[gemini-code-assist]

Code Review

This pull request removes NumPy and ml_dtypes as dependencies, standardizing on PyTorch for tensor operations across the CI scripts, documentation, and the code runner. The _to_torch utility was simplified by removing NumPy fallback logic. Feedback identifies redundant .cpu().contiguous() calls in simpler_setup/code_runner.py and suggests further consolidation of tensor conversion logic using the _to_torch helper.