Config max token default values mismatch

[akordowski]

I am wondering why some max token default values in config classes are set to 8191 and other to 8192. Is there any reason for this? Have seen this mismatch in other config classes too.

kernel-memory/extensions/OpenAI/OpenAIConfig.cs

Lines 50 to 64 in df05cdb

	/// <summary>
	/// The max number of tokens supported by the text model.
	/// </summary>
	public int TextModelMaxTokenTotal { get; set; } = 8192;
	/// <summary>
	/// Model used to embedding generation/
	/// </summary>
	public string EmbeddingModel { get; set; } = string.Empty;
	/// <summary>
	/// The max number of tokens supported by the embedding model.
	/// Default to OpenAI ADA2 settings.
	/// </summary>
	public int EmbeddingModelMaxTokenTotal { get; set; } = 8191;

[dluc]

It depends on the models:
8191 => see https://platform.openai.com/docs/guides/embeddings/embedding-models
8192 => see https://platform.openai.com/docs/models/gpt-4-turbo-and-gpt-4

[akordowski]

Thank you very much! At first glance it looked odd.