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, 04:45 UTC
[CI Failure] Fix OOM with test_oot_registration_embedding (#20144)
Signed-off-by: mgoin <mgoin64@gmail.com>
Michael Goin committed
Jun 27, 2025 at 12:21 UTC
71799fd005ca08c9c362e548945a3dde93790fec
1 file changed
+3
-1
tests/models/test_oot_registration.py
+3
-1
@@ -53,7 +53,9 @@ def test_oot_registration_embedding(
53
with monkeypatch.context() as m:
54
m.setenv("VLLM_PLUGINS", "register_dummy_model")
55
prompts = ["Hello, my name is", "The text does not matter"]
56
- llm = LLM(model=dummy_gemma2_embedding_path, load_format="dummy")
56
+ llm = LLM(model=dummy_gemma2_embedding_path,
57
+ load_format="dummy",
58
+ max_model_len=2048)
59
outputs = llm.embed(prompts)
60
61
for output in outputs: