fix: Default to port 8321 everywhere (#1734)

As titled, moved all instances of 5001 to 8321

llama_stack/templates/meta-reference-quantized-gpu/doc_template.md

@@ -67,7 +67,7 @@ You can do this via Conda (build code) or Docker which has a pre-built image.
 This method allows you to get started quickly without having to build the distribution code.
 
 ```bash
-LLAMA_STACK_PORT=5001
+LLAMA_STACK_PORT=8321
 docker run \
   -it \
   --pull always \

llama_stack/templates/meta-reference-quantized-gpu/meta_reference.py

@@ -100,7 +100,7 @@ def get_distribution_template() -> DistributionTemplate:
         },
         run_config_env_vars={
             "LLAMA_STACK_PORT": (
-                "5001",
+                "8321",
                 "Port for the Llama Stack distribution server",
             ),
             "INFERENCE_MODEL": (