GitLab

* ✨

 added an approach to use tokenizer_info endpoint from vllm
Signed-off-by: m-misiura <mmisiura@redhat.com>

* 🚧

 removed all auto-detection and tokenization logic from `LocalChatCompletion`

* pacify pre-commit

---------
Signed-off-by: m-misiura <mmisiura@redhat.com>
Co-authored-by: Baber <baber@hey.com>