fix: Handle Empty SSE Terminator in LM Studio/llama.cpp

[deathwish]

Description of Changes

Sporadically, reading from LM Studio or llama.cpp's OpenAI endpoint would hang with "streaming" status indicated despite inference having completed, likely due to MTP tearing. Google informed me that this was an expected behavior and the fix was recognizing an empty payload as an alternative terminator. I've been running this fix for several million tokens worth of generation and it's been rock solid, no hangs or early termination whatsoever.

Contributor License Agreement (CLA)

To accept your code, we legally need you to agree to our CLA so we can maintain the project's Business Source License (BSL) and future open-source transitions.

• By checking this box, I confirm that I have read and agree to the terms of the CLA.md in this repository. (To check the box, put an x between the brackets like this: [x])

[mlhher]

Thank you for your contribution. If everything looks and works fine this will be merged. Interesting though since I never ran into this.

[deathwish]

Awesome, thanks much!