Translating the grammar parser

SymbolIds is now SortedDictionary (although I'm not sure it really needs to be) because the test was failing due to expected value being in another order. The C++ data structure if SymbolIds is std::map<std::string, uint32_t> so the items are ordered by key.

Fixed those cublas paths again

Fixed the cublas linux paths again

fixed linux cublas paths in final step

Fixed linux cublas filenames

Linux cublas

Fixed cuda paths again

Fixed paths

* new github action to build deps (manually triggered)

Removed unnecessary parameters from some low level sampler methods

- Fixed multiple enumeration
 - Fixed newline penalisation

Two small improvements to the native sampling API

- Modified `llama_sample_token_mirostat` and `llama_sample_token_mirostat_v2` to take `ref float` instead of as a `float*`. Less pointers is always good.
 - Modified `llama_sample_repetition_penalty` and `llama_sample_frequency_and_presence_penalties` to take pointers instead of arrays. This allows the use non non allocating types (e.g. Span) instead of arrays
 - Modified higher level API to accept `Memory<int>` instead of `int[]`, which can be used to reduce allocations at call sites

ReSharper code warnings cleanup

ModelsParams record class

Improved Cloning

- Used those methods to add a `Clone` method to SafeLLamaContextHandle
 - Simplified `LLamaContext` by using the new methods
 - Sealed `LLamaContext` and `LLamaEmbedder`

Embedder Test