Added clarification between Qwen token count and word count

Removed tildes to prevent markdown strikethrough rendering issues

Clarified 550 Million token corpus size for CPT methodology verification

Added explicit CPU execution instructions and code example

Reverted to standard text completion code for Base/CPT model representation

Fixed code snippet to use apply_chat_template

Cleaned up model description (removed Lankabizz and hardware specs)

Upload README.md