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, 05:30 UTC
[CI] Fix distributed hybrid tests in CI (#26155)
Signed-off-by: Thomas Parnell <tpa@zurich.ibm.com>
Thomas Parnell committed
Oct 3, 2025 at 11:14 UTC
0e93ac0b3a18ebf8fa50f9263735fca5d69498ff
1 file changed
+2
-2
tests/models/language/generation/test_hybrid.py
+2
-2
@@ -240,12 +240,12 @@ def test_distributed_correctness(
240
num_logprobs: int,
241
) -> None:
242
with vllm_runner(model, tensor_parallel_size=1,
243
- max_num_seqs=2) as vllm_model:
243
+ max_num_seqs=MAX_NUM_SEQS) as vllm_model:
244
vllm_outputs_tp_1 = vllm_model.generate_greedy_logprobs(
245
example_prompts, max_tokens, num_logprobs)
246
247
with vllm_runner(model, tensor_parallel_size=2,
248
- max_num_seqs=2) as vllm_model:
248
+ max_num_seqs=MAX_NUM_SEQS) as vllm_model:
249
vllm_outputs_tp_2 = vllm_model.generate_greedy_logprobs(
250
example_prompts, max_tokens, num_logprobs)
251