[Core] Pipeline Parallel Support (#4412)

Signed-off-by: Muralidhar Andoorveedu <muralidhar.andoorveedu@centml.ai>

tests/tensorizer_loader/test_tensorizer.py

@@ -22,7 +22,7 @@ from vllm.model_executor.model_loader.tensorizer import (TensorizerConfig,
                                                          tensorize_vllm_model)
 
 from ..conftest import VllmRunner, cleanup
-from ..utils import RemoteOpenAIServer
+from ..utils import VLLM_PATH, RemoteOpenAIServer
 
 # yapf conflicts with isort for this docstring
 
@@ -220,6 +220,8 @@ def test_openai_apiserver_with_tensorizer(vllm_runner, tmp_path):
         json.dumps(model_loader_extra_config),
     ]
 
+    ray.init(runtime_env={"working_dir": VLLM_PATH})
+
     server = RemoteOpenAIServer(openai_args)
     print("Server ready.")