Swapped `StatelessExecutor` to use `llama_decode`!

- Growing more than double, if necessary

- Added `logits_i` argument to `Context.ApplyPenalty`
 - Added a new exception type for `llama_decode` return code

- using less `NativeHandle`
 - Using `StreamingTokenDecoder` instead of obsolete detokenize method

- Added a `DecodeAsync` overload which runs the work in a task
 - Replaced some `NativeHandle` usage in `BatchedDecoding` with higher level equivalents.
 - Made the `LLamaBatch` grow when token capacity is exceeded, removing the need to manage token capacity externally.

- Removed special case for exit
 - Added a wait at the end of batched decoding

- Renamed files and updated namespaces in Examples folder.
- Moved files from NewVersion folder to Examples folder.
- Removed TestRunner.cs file.
- Updated Runner.cs to include new examples.
- Update Program.cs to use the new Runner class instead of NewVersionTestRunner
- Update LLama.Examples namespace in Program.cs
- Update await NewVersionTestRunner.Run() in Program.cs to await Runner.Run()