A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 14, 05:30 UTC
[CI/Build] Fix tensorizer test for load_format change (#22583)
Signed-off-by: 22quinn <33176974+22quinn@users.noreply.github.com>
22quinn committed
Aug 10, 2025 at 19:30 UTC
b799f4b9ea8d15d62c4f4a97926b274561fd9492
3 files changed
+3
-4
.buildkite/test-pipeline.yaml
-1
@@ -426,7 +426,6 @@ steps:
426
427
- label: Tensorizer Test # 11min
428
mirror_hardwares: [amdexperimental]
429
- soft_fail: true
429
source_file_dependencies:
430
- vllm/model_executor/model_loader
431
- tests/tensorizer_loader
tests/entrypoints/openai/test_tensorizer_entrypoint.py
+1
-1
@@ -44,7 +44,7 @@ def model_uri(tmp_dir):
44
def tensorize_model_and_lora(tmp_dir, model_uri):
45
tensorizer_config = TensorizerConfig(tensorizer_uri=model_uri,
46
lora_dir=tmp_dir)
47
- args = EngineArgs(model=MODEL_NAME, device="cuda")
47
+ args = EngineArgs(model=MODEL_NAME)
48
49
tensorize_lora_adapter(LORA_PATH, tensorizer_config)
50
tensorize_vllm_model(args, tensorizer_config)
tests/tensorizer_loader/test_tensorizer.py
+2
-2
@@ -166,7 +166,7 @@ def test_load_without_tensorizer_load_format(vllm_runner, capfd, model_ref):
166
combined_output = out + err
167
assert ("ValueError: Model loader extra config "
168
"is not supported for load "
169
- "format LoadFormat.AUTO") in combined_output
169
+ "format auto") in combined_output
170
finally:
171
del model
172
gc.collect()
@@ -186,7 +186,7 @@ def test_raise_value_error_on_invalid_load_format(vllm_runner, capfd,
186
187
combined_output = out + err
188
assert ("ValueError: Model loader extra config is not supported "
189
- "for load format LoadFormat.SAFETENSORS") in combined_output
189
+ "for load format safetensors") in combined_output
190
finally:
191
del model
192
gc.collect()