Get up and running with Kimi, GLM, MiniMax, DeepSeek, gpt-oss, Qwen, Gemma and other models.
Reset NumCtx.
Richard Lyons9352eeb75253
diff --git a/server/sched.go b/server/sched.goindex 4d9c02969..3fe6d7fc9 100644--- a/server/sched.go+++ b/server/sched.go@@ -736,6 +736,7 @@ func pickBestFullFitByLibrary(req *LlmRequest, ggml *llm.GGML, gpus gpu.GpuInfoLfunc pickBestPartialFitByLibrary(req *LlmRequest, ggml *llm.GGML, gpus gpu.GpuInfoList, numParallel *int) gpu.GpuInfoList {if *numParallel <= 0 {*numParallel = 1+ req.opts.NumCtx = req.origNumCtx}byLibrary := gpus.ByLibrary()if len(byLibrary) <= 1 {