Unify the chat api for all inference servers #218
Description
What would you like to be added:
The chat api:
- POST /v1/chat/completions
The completion api:
- POST /v1/completions
Why is this needed:
Reduce the operating cost for users. If we can unify the post content, this could be better.
Completion requirements:
This enhancement requires the following artifacts:
- Design doc
- API change
- Docs update
The artifacts should be linked in subsequent comments.