fix: do not override the entrypoint when running container

Since https://github.com/meta-llama/llama-stack/pull/2005 the run
configuration is embedded into the container image itself and the
entrypoint is correctly configured during the container image build
process. We don't need to override the container image entrypoint
anymore.

Signed-off-by: Sébastien Han <seb@redhat.com>

llama_stack/distribution/start_stack.sh

@@ -152,12 +152,8 @@ elif [[ "$env_type" == "container" ]]; then
     $CONTAINER_BINARY run $CONTAINER_OPTS -it \
     -p $port:$port \
     $env_vars \
-    -v "$yaml_config:/app/config.yaml" \
     $mounts \
     --env LLAMA_STACK_PORT=$port \
-    --entrypoint python \
     $container_image:$version_tag \
-    -m llama_stack.distribution.server.server \
-    --yaml-config /app/config.yaml \
     $other_args
 fi