From 8a2a723d4b77eb64ee4c7cf8e1168adb4a664ea9 Mon Sep 17 00:00:00 2001 From: Hyperling Date: Sat, 17 Oct 2026 12:01:57 -0700 Subject: [PATCH] llm, continue dropping keep alive since tools which use LLMs correctly manage this themselves. --- tasks/workstation/linux/software/llm.yml | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tasks/workstation/linux/software/llm.yml b/tasks/workstation/linux/software/llm.yml index 34beaf9..bc9c847 100644 --- a/tasks/workstation/linux/software/llm.yml +++ b/tasks/workstation/linux/software/llm.yml @@ -256,7 +256,7 @@ export OLLAMA_REQUEST_TIMEOUT=3600 # default 300 (unconfirmed, not in the list of --help vars) ### Resource Usage ### - export OLLAMA_KEEP_ALIVE=2h # default is 5m + export OLLAMA_KEEP_ALIVE=20m # default is 5m export OLLAMA_MAX_LOADED_MODELS={{ llm_concurrency }} # default is 3 if using CPUs, or 3 * NumGPUs export OLLAMA_NUM_PARALLEL={{ llm_parallel }} # default is 1 and applies per-model (unconfirmed) export OLLAMA_CONTEXT_LENGTH={{ llm_context }} # default is 4096, https://docs.ollama.com/faq#how-can-i-specify-the-context-window-size