[CI/Build][AMD] Use float16 in test_reset_prefix_cache_e2e to avoid accuracy issues (#29997)

Signed-off-by: Randall Smith <ransmith@amd.com> Co-authored-by: Randall Smith <ransmith@amd.com>
2025-12-05 02:42:25 -06:00
parent 6038b1b04b
commit feecba09af
1 changed files with 1 additions and 0 deletions
--- a/tests/v1/core/test_reset_prefix_cache_e2e.py
+++ b/tests/v1/core/test_reset_prefix_cache_e2e.py
@@ -21,6 +21,7 @@ def test_reset_prefix_cache_e2e(monkeypatch):
        max_num_batched_tokens=32,
        max_model_len=2048,
        compilation_config={"mode": 0},
+        dtype="float16",
    )
    engine = LLMEngine.from_engine_args(engine_args)
    sampling_params = SamplingParams(