fix non-AVX CPU detection

[cebtenzzre]

Key Changes

• Fix a missing defined(_M_X64) that caused the "you don't have AVX" screen to not appear on Windows
• Refactor CPU feature detection so that can't happen anymore
• Improve the way we report this to the bindings

Chat

This message can now be shown on Windows. This screenshot is only an example, I haven't actually tested this on a Windows machine without AVX:

Bindings

Before (exception is generic and misleading, actual error is hidden above the traceback):

LLModel ERROR: CPU does not support AVX
Traceback (most recent call last):
  File "/home/jared/src/own/gpt4all-scripts/test_any.py", line 18, in <module>
    main()
  File "/home/jared/src/own/gpt4all-scripts/test_any.py", line 10, in main
    x = GPT4All(model_name=path.name, model_path=path.parent, allow_download=False, device="Tesla P40")
        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/jared/src/forks/gpt4all/gpt4all-bindings/python/gpt4all/gpt4all.py", line 132, in __init__
    self.model = _pyllmodel.LLModel(self.config["path"], n_ctx, ngl)
                 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/jared/src/forks/gpt4all/gpt4all-bindings/python/gpt4all/_pyllmodel.py", line 188, in __init__
    raise ValueError(f"Unable to instantiate model: {'null' if s is None else s.decode()}")
ValueError: Unable to instantiate model: Model format not supported (no matching implementation found)

After:

Traceback (most recent call last):
  File "/home/jared/src/own/gpt4all-scripts/test_any.py", line 18, in <module>
    main()
  File "/home/jared/src/own/gpt4all-scripts/test_any.py", line 10, in main
    x = GPT4All(model_name=path.name, model_path=path.parent, allow_download=False, device="Tesla P40")
        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/jared/src/forks/gpt4all/gpt4all-bindings/python/gpt4all/gpt4all.py", line 132, in __init__
    self.model = _pyllmodel.LLModel(self.config["path"], n_ctx, ngl)
                 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/jared/src/forks/gpt4all/gpt4all-bindings/python/gpt4all/_pyllmodel.py", line 188, in __init__
    raise RuntimeError(f"Unable to instantiate model: {'null' if s is None else s.decode()}")
RuntimeError: Unable to instantiate model: CPU does not support AVX