[CI/Build] Fix tensorizer test for load_format change (#22583)

Signed-off-by: 22quinn <[email protected]>

Author: 22quinn

.buildkite/test-pipeline.yaml

@@ -426,7 +426,6 @@ steps:
 
 - label: Tensorizer Test # 11min
   mirror_hardwares: [amdexperimental]
-  soft_fail: true
   source_file_dependencies:
   - vllm/model_executor/model_loader
   - tests/tensorizer_loader

tests/entrypoints/openai/test_tensorizer_entrypoint.py

@@ -44,7 +44,7 @@ def model_uri(tmp_dir):
 def tensorize_model_and_lora(tmp_dir, model_uri):
     tensorizer_config = TensorizerConfig(tensorizer_uri=model_uri,
                                          lora_dir=tmp_dir)
-    args = EngineArgs(model=MODEL_NAME, device="cuda")
+    args = EngineArgs(model=MODEL_NAME)
 
     tensorize_lora_adapter(LORA_PATH, tensorizer_config)
     tensorize_vllm_model(args, tensorizer_config)

tests/tensorizer_loader/test_tensorizer.py

@@ -166,7 +166,7 @@ def test_load_without_tensorizer_load_format(vllm_runner, capfd, model_ref):
         combined_output = out + err
         assert ("ValueError: Model loader extra config "
                 "is not supported for load "
-                "format LoadFormat.AUTO") in combined_output
+                "format auto") in combined_output
     finally:
         del model
         gc.collect()
@@ -186,7 +186,7 @@ def test_raise_value_error_on_invalid_load_format(vllm_runner, capfd,
 
         combined_output = out + err
         assert ("ValueError: Model loader extra config is not supported "
-                "for load format LoadFormat.SAFETENSORS") in combined_output
+                "for load format safetensors") in combined_output
     finally:
         del model
         gc.collect()