From 822a0e36eb15457fb04c1a42d34cf0db938a9789 Mon Sep 17 00:00:00 2001 From: Jeffrey Morgan Date: Fri, 28 Jul 2023 10:56:21 -0400 Subject: [PATCH] lower batch size to 512 --- api/types.go | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/api/types.go b/api/types.go index e8f91270..2e3b0578 100644 --- a/api/types.go +++ b/api/types.go @@ -190,7 +190,7 @@ func DefaultOptions() Options { UseNUMA: false, NumCtx: 2048, - NumBatch: 1024, + NumBatch: 512, NumGPU: 1, NumGQA: 1, LowVRAM: false,