From d8b3e09fb7bac44eab8ad03e1394eb10196ed4c0 Mon Sep 17 00:00:00 2001 From: jmorganca Date: Sat, 8 Jun 2024 22:55:22 -0700 Subject: [PATCH] llm: enable flash attention by default --- envconfig/config.go | 1 + 1 file changed, 1 insertion(+) diff --git a/envconfig/config.go b/envconfig/config.go index ae4e99391..a2d20f303 100644 --- a/envconfig/config.go +++ b/envconfig/config.go @@ -90,6 +90,7 @@ func init() { NumParallel = 1 MaxRunners = 1 MaxQueuedRequests = 512 + FlashAttention = true LoadConfig() }