Bart Czernicki
Bart Czernicki
I think having some advanced scenarios/architectures that someone can just jump to or see some demos would be nice (This way you have a "forward looking", this is all the...
This database concept uses SQL Server or Azure SQL Database engine features that allow you to process vectors using OpenAI embeddings and GPTx models. All of the code shown below...
The OpenAI.NET library has this working and routing through the MaxOutputTokenCount. Can this be fixed or applied with parity as this block o1 paramater settings.
You will need the new Azure.AI.Projects package (beta 9). It appears the Azure.AI.Projects has changed the AIAgent object to PersistentAgent and it breaks the SK integration with Azure AI Foundry.
Bump...pretty much can't use Interactive Notebooks without .NET 10 support.
Bump...pretty much can't use Interactive Notebooks without .NET 10 support.
This is largely temporary, as you need to go back to when Microsoft.Extensions.AI was on .NET 9 and you will be missing new features: https://github.com/dotnet/interactive/issues/4086#issuecomment-3587477173
> is it fixed? Yes, reasoning models now have the reasoning effort controls supported.
@vovanb You need todo something like this: ChatCompletionOptions options = new() { MaxOutputTokenCount = 200 }; // For newer models (GPT-5, o4, etc.) options.SetNewMaxCompletionTokensPropertyEnabled(true);