[TPU][Bugfix] fix OOM issue in CI test (#21550)

Signed-off-by: Chengji Yao <chengjiyao@google.com>

[TPU][Bugfix] fix OOM issue in CI test (#21550)
Signed-off-by: Chengji Yao <chengjiyao@google.com>
40d86ee4 · Chengji Yao · GitHub · 85d051f0 · 40d86ee4
Unverified Commit 40d86ee4 authored Jul 24, 2025 by Chengji Yao Committed by GitHub Jul 24, 2025
Hide whitespace changes
Inline Side-by-side

Showing with 1 addition and 1 deletion

tests/v1/tpu/test_basic.py tests/v1/tpu/test_basic.py +1 -1

No files found.
--- a/tests/v1/tpu/test_basic.py
+++ b/tests/v1/tpu/test_basic.py
@@ -59,7 +59,7 @@ def test_basic(
                # actually test chunked prompt
                max_num_batched_tokens=1024,
                max_model_len=8192,
-                gpu_memory_utilization=0.7,
+                gpu_memory_utilization=0.95,
                max_num_seqs=max_num_seqs,
                tensor_parallel_size=tensor_parallel_size) as vllm_model:
            vllm_outputs = vllm_model.generate_greedy(example_prompts,