llama_sample_classifier_free_guidance

Logger Comments

Improved `NativeApi` file a bit

- Added some more comments
 - Modified `llama_tokenize` to not allocate
 - Modified `llama_tokenize_native` to take a pointer instead of an array, allowing use with no allocations
 - Removed GgmlInitParams (not used)

Minor quantizer improvements

Add native logging output

- Rewritten `LLamaFtype` parsing to support any substring which uniquely matches a single enum variant

Fix serialization error due to NaN

Fixed `ToLlamaContextParams` using the wrong parameter for `use_mmap`

Utils Cleanup

Access to IModelParamsExtensions

fix: antiprompt does not work in stateless executor.

- Added disposal check into `SafeLLamaContextHandle`

- Also refactored it to return an `int[]` instead of an `IEnumerable<int>`, solving the "multiple enumeration" problems at the source!

Improved Tensor Splits

Cleaned up multiple enumeration in FixedSizeQueue

ModelParams & InferenceParams abstractions

feat: update the llama backends.

Expose some native classes