diff --git a/serving.proto b/serving.proto index e2861bd..a16dde1 100755 --- a/serving.proto +++ b/serving.proto @@ -122,6 +122,10 @@ message SummarizationRequest { // Higher values penalize new tokens based on whether they appear in the text so far, // increasing the model's likelihood to talk about new topics. optional float presence_penalty = 20; + // This feature is in Beta. If specified, the LLM will make a best effort to sample deterministically, + // such that repeated requests with the same seed and parameters should return the same result. + // Note:- Seed value is only used when temperature is set to 0. + optional uint32 seed = 25; } ModelParams model_params = 215; @@ -216,8 +220,8 @@ message QueryRequest { message RerankingConfig { // Which reranking model to use if reranking. Currently, the only IDs // available are: - // - 272725717, HuggingFace Open Source x-attentional reranker // - 272725718, Maximum Marginal Relevance Reranker + // - 272725719, Vectara Multilingual Reranker v1 Reranker uint32 reranker_id = 5; // Reranker-specific parameters. The numbering starts from 100, and moves @@ -322,7 +326,6 @@ message Summary { repeated Status status = 1000; // Populated for streaming requests only. int32 future_id = 1010; - } // A document part that matched a query.