Make `LLamaKvCacheView` Safe

Llama Text Templater

Fix cublas build action

* ci: add windows benchmark test.

* ci: add benchmark test.

- Checking indices
 - Returning span of correct length
 - Hiding native methods
 - Hiding native types

- Returning template for chaining method calls
 - Returning a `TextMessage` object instead of a tuple

- Fixed `llama_chat_apply_template` method (wrong entrypoint, couldn't handle null model)

Remove `Conversation.Prompt(String)`

Implement context shifting in executor base

SemanticKernel: Correcting non-standard way of working with PromptExecutionSettings

Optimize Semantic Kernel LLamaSharpChatCompletion when running with StatefulExecutorBase models

Optional IHistoryTransform added to ChatSession.InitializeSessionFromHistoryAsync

Optional IHistoryTransform added to ChatSession.InitializeSessionFromHistoryAsync

ChatSession.InitializeSessionFromHistoryAsync now accepts IHistoryTransform as an optional parameter.

ci: add workflow to check the spellings.

Add LLaMA3 chat session example.

LLava Async Loading

- Fixed checking for invalid handles in `clip_model_load`

Interruptible Async Model Loading With Progress Monitoring

can be used downstream to post-process the messages based on the requested format

- Async loading supports cancellation through a `CancellationToken`. If loading is cancelled an `OperationCanceledException` is thrown.  If it fails for another reason a `LoadWeightsFailedException` is thrown.
 - Updated examples to use `LoadFromFileAsync`

Fix typo in issue templates.

Add issue templates.

Removed Unnecessary Constructor From Safe Handles

Empty batch check

Slightly Safer Quantize Params

- Passing model path, instead of a message, to `LoadWeightsFailedException` constructor

- Returning SafeLLamaGrammarHandle directly from `llama_grammar_init` and `llama_grammar_copy`