[ROCm][CI] Add large_gpu_mark to test_max_tokens_none for ROCm (#37717)
Signed-off-by: Andreas Karatzas <akaratza@amd.com>
This commit is contained in:
@@ -12,6 +12,7 @@ import gc
|
|||||||
import pytest
|
import pytest
|
||||||
import torch
|
import torch
|
||||||
|
|
||||||
|
from tests.utils import large_gpu_mark
|
||||||
from vllm import LLM, SamplingParams
|
from vllm import LLM, SamplingParams
|
||||||
from vllm.platforms import current_platform
|
from vllm.platforms import current_platform
|
||||||
|
|
||||||
@@ -32,10 +33,21 @@ def test_duplicated_ignored_sequence_group():
|
|||||||
assert len(prompts) == len(outputs)
|
assert len(prompts) == len(outputs)
|
||||||
|
|
||||||
|
|
||||||
def test_max_tokens_none():
|
@pytest.mark.parametrize(
|
||||||
|
"model",
|
||||||
|
[
|
||||||
|
pytest.param(
|
||||||
|
"distilbert/distilgpt2",
|
||||||
|
marks=[
|
||||||
|
*([large_gpu_mark(min_gb=80)] if current_platform.is_rocm() else []),
|
||||||
|
],
|
||||||
|
),
|
||||||
|
],
|
||||||
|
)
|
||||||
|
def test_max_tokens_none(model):
|
||||||
sampling_params = SamplingParams(temperature=0.01, top_p=0.1, max_tokens=None)
|
sampling_params = SamplingParams(temperature=0.01, top_p=0.1, max_tokens=None)
|
||||||
llm = LLM(
|
llm = LLM(
|
||||||
model="distilbert/distilgpt2",
|
model=model,
|
||||||
max_num_batched_tokens=4096,
|
max_num_batched_tokens=4096,
|
||||||
tensor_parallel_size=1,
|
tensor_parallel_size=1,
|
||||||
)
|
)
|
||||||
|
|||||||
Reference in New Issue
Block a user