Update start.sh

It will give boost upto 3x times. As docker container uses 2 cpu cores, we get best performance when threads are same as CPU count. By default it used 8 process.

start.sh

@@ -2,7 +2,7 @@
 
 # Start llama-server in background
 cd /llama.cpp/build
-./bin/llama-server --host 0.0.0.0 --port 8080 --model /models/model.q8_0.gguf --ctx-size 32768 &
+./bin/llama-server --host 0.0.0.0 --port 8080 --model /models/model.q8_0.gguf --ctx-size 32768 --threads 2  &
 
 # Wait for server to initialize
 echo "Waiting for server to start..."