ModelSpec not working as expected after recent commits #4213
-
What happened?I made no changes to my code except updating to recent commits, and my modelSpec with custom OpenAI endpoint has stopped working. My gptPlugins model spec, that uses the OPENAI_REVERSE_PROXY base, still works fine, but my custom openAI endpoint returns a modelspec mismatch, despite having selected the Model. Steps to Reproduce
What browsers are you seeing the problem on?Safari, Microsoft Edge, Mobile (iOS), Mobile (Android) Relevant log outputNo response ScreenshotsCode of Conduct
|
Beta Was this translation helpful? Give feedback.
Replies: 2 comments 4 replies
-
Try testing it after this PR is merged. It makes more sense to apply the configured preset, regardless of what the client sent, instead of checking for a mismatch. |
Beta Was this translation helpful? Give feedback.
-
I had the same issue with Azure OpenAI, and I got it fixed by setting enforce to false. I'm not 100% sure what that option does, if all the interface parameters are turned off anyways? It's ok right to leave enforce to false then? |
Beta Was this translation helpful? Give feedback.
Try testing it after this PR is merged.
#4214
It makes more sense to apply the configured preset, regardless of what the client sent, instead of checking for a mismatch.