mirror of
https://github.com/microsoft/semantic-kernel.git
synced 2026-10-12 02:19:06 +00:00
Closed #2417 in favour of this PR which also includes @relango s commits. This PR updates for full support for ADR 008 (https://github.com/microsoft/semantic-kernel/blob/main/docs/decisions/0008-support-generic-llm-request-settings.md), which is the ADR in this area. This PR does the following: - Updates the code to support ADR 008. - Apply code formatting needed on these - If you dont want do do this PR via your branch I can PR it directly and close this PR. Now you can add serviceId and model to invocation, i.e: `` Mono<SKContext> result = function.invokeAsync("time travel to dinosaur age", SKBuilders .completionRequestSettings() .serviceId("text-davinci-003") .build()); `` function config, i.e: ``` var function = kernel. getSemanticFunctionBuilder() .withPromptTemplate("tell a joke about {{$input}}") .withRequestSettings( SKBuilders.completionRequestSettings() .temperature(0.4) .topP(1) .maxTokens(100) .serviceId("text-davinci-003") .build()) .build(); ``` Or config file ``` { "schema": 1, "description": "Sample plugin with a model and service id", "type": "completion", "completion": { "max_tokens": 150, "temperature": 0.9, "top_p": 0.0, "presence_penalty": 0.6, "frequency_penalty": 0.0, "service_id": "text-davinci-003", "model_id": "text-davinci-003" } } ``` --------- Co-authored-by: Rajasekar Elango <[email protected]>