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, 23:45 UTC
[TPU][TEST] HF_HUB_DISABLE_XET=1 the test 3. (#21539)
Signed-off-by: Qiliang Cui <derrhein@gmail.com>
QiliangCui committed
Jul 24, 2025 at 15:33 UTC
07d80d7b0eeed29564345337258600a2cf22481d
2 files changed
+1
-4
.buildkite/scripts/hardware_ci/run-tpu-v1-test.sh
+1
-1
@@ -135,7 +135,7 @@ run_and_track_test 1 "test_compilation.py" \
135
run_and_track_test 2 "test_basic.py" \
136
"python3 -m pytest -s -v /workspace/vllm/tests/v1/tpu/test_basic.py"
137
run_and_track_test 3 "test_accuracy.py::test_lm_eval_accuracy_v1_engine" \
138
- "python3 -m pytest -s -v /workspace/vllm/tests/entrypoints/llm/test_accuracy.py::test_lm_eval_accuracy_v1_engine"
138
+ "HF_HUB_DISABLE_XET=1 python3 -m pytest -s -v /workspace/vllm/tests/entrypoints/llm/test_accuracy.py::test_lm_eval_accuracy_v1_engine"
139
run_and_track_test 4 "test_quantization_accuracy.py" \
140
"python3 -m pytest -s -v /workspace/vllm/tests/tpu/test_quantization_accuracy.py"
141
run_and_track_test 5 "examples/offline_inference/tpu.py" \
tests/entrypoints/llm/test_accuracy.py
-3
@@ -73,9 +73,6 @@ def test_lm_eval_accuracy_v1_engine(model, monkeypatch: pytest.MonkeyPatch):
73
if current_platform.is_tpu():
74
# Limit compilation time for TPU V1
75
76
- # xet doesn't work well for both Qwen/Qwen3-1.7B and
77
- # google/gemma-3-1b-it
78
- m.setenv("HF_HUB_DISABLE_XET", "1")
76
more_args = "max_model_len=2048,max_num_seqs=64"
77
78
# Add TP test (if provided)