From f9b4e0204784ef76deb1aeeca98c7e7f777b2155 Mon Sep 17 00:00:00 2001
From: Trevor SANDY <trevor.sandy@gmail.com>
Date: Wed, 21 Jan 2026 10:43:31 +0100
Subject: [PATCH] Change Ollama and LLaMA.cpp context size in .env to 4096 from
 8192

---
 .env.example | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/.env.example b/.env.example
index efdf1da..6089df6 100644
--- a/.env.example
+++ b/.env.example
@@ -213,7 +213,7 @@ OLLAMA_PORT=11434
 OLLAMA_HOST=host.docker.internal:11434
 
 # Tuning
-OLLAMA_CONTEXT_LENGTH=8192
+OLLAMA_CONTEXT_LENGTH=4096
 OLLAMA_FLASH_ATTENTION=1
 OLLAMA_KV_CACHE_TYPE=q4_0
 OLLAMA_MAX_LOADED_MODELS=2
@@ -271,7 +271,7 @@ LLAMACPP_MODEL_PATH=${LLAMACPP_MODELS_DIR}/${LLAMACPP_MODEL_NAME}
 LLAMA_ARG_HF_REPO=${LLAMACPP_MODEL_GEMMA_ID}
 
 # Tuning
-LLAMA_ARG_CTX_SIZE=8192
+LLAMA_ARG_CTX_SIZE=4096
 LLAMA_ARG_FLASH_ATTN=1
 LLAMA_ARG_N_GPU_LAYERS=0
 LLAMA_ARG_THREADS=4