MinP Sampler

- Cleaned up comments in implementations of `IInferenceParams`
 - Removed default values for all parameters in `LLamaContext.Sample` - they're never used and probably _shouldn't_ ever be used

This reverts commit 091b8d58b3502a99b3bfbec9db457c92cc736beb, reversing
changes made to 9b2ca9cf8e.

This reverts commit b4b3ea9d99.

Prevent duplication of user prompts / chat history in ChatSession.

Just merge cuda and avx detection and change layout.

Delete Backend.Metal because is not needed anymore.
Do not include .metal in x86_64 binaries

The way ChatSession.ChatAsync was using the provided methods
from a IHistoryTransform interface implementation created unexpected
duplication of the chat history messages. It also prevented loading
previous history into the current session.

Create different paths to different MacOS platforms.
Dynamically load the right library

Add the MacOS binary from the same run

Should fix the extreme speed loss.

- Removed all `record struct` uses in native code
 - Removed usage of `readonly` in native structs

Minor fix:
 - Added sequential layout to `LLamaModelQuantizeParams`

had to update binary to `b1426`