From 4c79f8a0ab7f2c36c03a338b1b72013570fe7647 Mon Sep 17 00:00:00 2001 From: "Wang, Kai Lawrence" Date: Fri, 20 Dec 2024 14:56:16 +0800 Subject: [PATCH] Set topK=1 for faqgenfixed Signed-off-by: Wang, Kai Lawrence --- evals/benchmark/stresscli/locust/faqgenfixed.py | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/evals/benchmark/stresscli/locust/faqgenfixed.py b/evals/benchmark/stresscli/locust/faqgenfixed.py index a121c555..2036c64c 100644 --- a/evals/benchmark/stresscli/locust/faqgenfixed.py +++ b/evals/benchmark/stresscli/locust/faqgenfixed.py @@ -9,7 +9,11 @@ def getUrl(): def getReqData(): - return {"messages": "Text Embeddings Inference (TEI) is a toolkit for deploying and serving open source text embeddings and sequence classification models. TEI enables high-performance extraction for the most popular models, including FlagEmbedding, Ember, GTE and E6.", "max_tokens": 128} + return { + "messages": "Text Embeddings Inference (TEI) is a toolkit for deploying and serving open source text embeddings and sequence classification models. TEI enables high-performance extraction for the most popular models, including FlagEmbedding, Ember, GTE and E6.", + "max_tokens": 128, + "top_k": 1, + } def respStatics(environment, reqData, respData):