Added int4 compression to LLM chatbot notebook (#1428)

* Added compression code

* Description changes

* Tweaks

* Disabled int4 compression for red-pajama model

* Tweak

* Readme tweak

* Made mpt model a default one

* Fix linter issues

* Addressed comments

* Tweaks

* Changed int4/int8 to 4/8 bit data types

* Included int8 in a note about dgpu

* Tweaked notes

* Added device

* Added selecting model to run

* Tweaked note