diff --git a/llama_stack/distribution/distribution.py b/llama_stack/distribution/distribution.py index 87d77b5ed..53d544471 100644 --- a/llama_stack/distribution/distribution.py +++ b/llama_stack/distribution/distribution.py @@ -35,7 +35,6 @@ def builtin_automatically_routed_apis() -> List[AutoRoutedApiInfo]: routing_table_api=Api.memory_banks, router_api=Api.memory, ), - # TODO: re-enable once we have proper checking on builtin routing: Issue #297 AutoRoutedApiInfo( routing_table_api=Api.datasets, router_api=Api.datasetio, diff --git a/llama_stack/distribution/server/server.py b/llama_stack/distribution/server/server.py index 178fac070..185c89e7e 100644 --- a/llama_stack/distribution/server/server.py +++ b/llama_stack/distribution/server/server.py @@ -288,6 +288,7 @@ def main( apis_to_serve = set(impls.keys()) for inf in builtin_automatically_routed_apis(): + # if we do not serve the corresponding router API, we should not serve the routing table API if inf.router_api.value not in apis_to_serve: continue apis_to_serve.add(inf.routing_table_api.value) diff --git a/tests/examples/local-run.yaml b/tests/examples/local-run.yaml index 128008117..e12f6e852 100644 --- a/tests/examples/local-run.yaml +++ b/tests/examples/local-run.yaml @@ -13,18 +13,14 @@ apis: - safety providers: inference: - - provider_id: tgi0 - provider_type: remote::tgi + - provider_id: meta-reference + provider_type: meta-reference config: - url: http://127.0.0.1:5009 - # - provider_id: meta-reference - # provider_type: meta-reference - # config: - # model: Llama3.1-8B-Instruct - # quantization: null - # torch_seed: null - # max_seq_len: 4096 - # max_batch_size: 1 + model: Llama3.1-8B-Instruct + quantization: null + torch_seed: null + max_seq_len: 4096 + max_batch_size: 1 safety: - provider_id: meta-reference provider_type: meta-reference