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 15, 21:00 UTC
[V1][Easy] Add empty allowed_token_ids in the v1 sampler test (#14308)
Signed-off-by: Lu Fang <lufang@fb.com>
Lu Fang committed
Mar 5, 2025 at 13:41 UTC
53ea6ad8304f5d16bab97527ea1f4a19cfb25588
1 file changed
+4
tests/v1/sample/test_sampling_params_e2e.py
+4
@@ -113,6 +113,10 @@ def test_allowed_token_ids(model):
113
PROMPT, SamplingParams(allowed_token_ids=allowed_token_ids))
114
assert output[0].outputs[0].token_ids[-1] == TOKEN_ID
115
116
+ # Reject empty allowed_token_ids.
117
+ with pytest.raises(ValueError):
118
+ _ = model.generate(PROMPT, SamplingParams(allowed_token_ids=[]))
119
+
120
# Reject negative token id.
121
with pytest.raises(ValueError):
122
_ = model.generate(PROMPT, SamplingParams(allowed_token_ids=[-1]))