Fixing some minor typos and spelling errors which should not affect functionality but improve the overall quality of documentation.

link in table of contents doesn't work without this change or changing the link fragment to `#single-gpu-`

avaiable -> available

Remove missing link

Moved the inference.md to docs to have the docs in one folder.

Created a Readme for the inference folder

Updated relevant docs to reflect new inference.md link.

Doc requirements update

remove unneeded imports for llama_finetuning.py