norohind/ollama
1
0
Fork 0
forked from third-party-mirrors/ollama
Code Pull Requests Activity

lower batch size to 512

Browse Source
This commit is contained in:
Jeffrey Morgan 2023-07-28 10:56:21 -04:00
parent 18b6b601ad
commit 822a0e36eb
1 changed files with 1 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
api/types.go
View File

@ -190,7 +190,7 @@ func DefaultOptions() Options {
UseNUMA: false,
NumCtx: 2048,
NumBatch: 1024,
NumBatch: 512,
NumGPU: 1,
NumGQA: 1,
LowVRAM: false,