mirrors/ollama-for-amd
1
0
Fork 0
mirror of https://github.com/likelovewant/ollama-for-amd.git synced 2025-12-21 22:33:56 +00:00
Code Issues Packages Projects Releases Wiki Activity

merge upstream and fix conflicts

Browse Source
This commit is contained in:
likelovewant
2024-12-12 13:17:55 +08:00
parent a0caaa2bc8
commit a64347c6d4
383 changed files with 59815 additions and 43978 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
integration/concurrency_test.go
View File

@@ -207,7 +207,7 @@ func TestMultiModelStress(t *testing.T) {
chosenModels = mediumModels
// default:
// slog.Info("selecting large models")
// chosenModels = largModels
// chosenModels = largeModels
}
req, resp := GenerateRequests()
@@ -232,7 +232,7 @@ func TestMultiModelStress(t *testing.T) {
var wg sync.WaitGroup
consumed := uint64(256 * format.MebiByte) // Assume some baseline usage
for i := 0; i < len(req); i++ {
// Always get at least 2 models, but dont' overshoot VRAM too much or we'll take too long
// Always get at least 2 models, but don't overshoot VRAM too much or we'll take too long
if i > 1 && consumed > maxVram {
slog.Info("achieved target vram exhaustion", "count", i, "vram", format.HumanBytes2(maxVram), "models", format.HumanBytes2(consumed))
break