REST completions API inconsistency

[samuel100]

In the REST API for inference (completions) tokens usage data is missing and the model name. Also, the casing is inconsistent (some are pascal , some are camel case).

@waldekmastykarz

AB#66026

[jdubois]

I contributed to the AI Foundry bug bash and I reported that issue of the model name being null: https://github.com/ai-platform-microsoft/WindowsAIFoundry/issues/124

I just tested with version 0.3.9266.20402 and I'm still having the issue of the model being null: as this is required by the OpenAI SDK (at least its Java version), that's a blocker issue for me.

cc @agoncal as he has the same issue as me

[jdubois]

I'd love to fix this by myself, but the server code doesn't seem to be in this repository, could you point me to where this is?

[jdubois]

Hello, this is blocking me from using the OpenAI Java SDK with Foundry Local, as this is a required argument, so this is very important for me -> how can I help fix the issue with the null model name?

[jdubois]

Hello, it's been 3 months and this bug doesn't seem to have been prioritized, how can I help?

[jdubois]

Hey @leestott we discussed this at the beginning of the summer, can you help on this? I'll be happy to fix it myself if I can get access to the repo.

[RichardC64]

Hummm @jdubois : SDK are Open Source but FoundryLocal is Microsoft Software Licence. Perharps it's the reason why there is no source code for it. That's strange.

[jdubois]

Hey, it's been nearly 4 months, any progress? I would still be happy to contribute.

[jdubois]

I just tested it again and this is fixed on my side! I'll stop following this issue, thank you!