xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2025-12-11 04:54:59 +08:00
Code Issues Packages Projects Releases Wiki Activity

[V1] TPU - CI/CD use smaller model (#15054)

Browse Source 
Signed-off-by: Alexander Matveev <amatveev@redhat.com>
This commit is contained in:
Alexander Matveev 2025-03-18 17:39:21 -04:00 committed by GitHub
parent 99abb8b650
commit 72a8639b68
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
2 changed files with 13 additions and 12 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

18
.buildkite/run-tpu-v1-test.sh
View File

@ -20,17 +20,17 @@ docker run --privileged --net host --shm-size=16G -it \
&& python3 -m pip install pytest \ && python3 -m pip install pytest \
&& python3 -m pip install lm_eval[api]==0.4.4 \ && python3 -m pip install lm_eval[api]==0.4.4 \
&& echo TEST_1 \ && echo TEST_1 \
&& VLLM_USE_V1=1 pytest -v -s /workspace/vllm/tests/v1/tpu/test_basic.py \
&& echo TEST_2 \
&& VLLM_USE_V1=1 pytest -v -s /workspace/vllm/tests/entrypoints/llm/test_accuracy.py::test_lm_eval_accuracy_v1_engine \
&& echo TEST_3 \
&& VLLM_USE_V1=1 pytest -s -v /workspace/vllm/tests/tpu/test_quantization_accuracy.py \
&& echo TEST_4 \
&& VLLM_USE_V1=1 python3 /workspace/vllm/examples/offline_inference/tpu.py" \
&& echo TEST_5 \
&& VLLM_USE_V1=1 python3 /workspace/vllm/tests/tpu/test_compilation.py \ && VLLM_USE_V1=1 python3 /workspace/vllm/tests/tpu/test_compilation.py \
&& echo TEST_2 \
&& VLLM_USE_V1=1 pytest -v -s /workspace/vllm/tests/v1/tpu/test_basic.py \
&& echo TEST_3 \
&& VLLM_USE_V1=1 pytest -v -s /workspace/vllm/tests/entrypoints/llm/test_accuracy.py::test_lm_eval_accuracy_v1_engine \
&& echo TEST_4 \
&& VLLM_USE_V1=1 pytest -s -v /workspace/vllm/tests/tpu/test_quantization_accuracy.py \
&& echo TEST_5 \
&& VLLM_USE_V1=1 python3 /workspace/vllm/examples/offline_inference/tpu.py" \
# TODO: Fix these tests # TODO: This test fails because it uses RANDOM_SEED sampling
# && VLLM_USE_V1=1 pytest -v -s /workspace/vllm/tests/tpu/test_custom_dispatcher.py \ # && VLLM_USE_V1=1 pytest -v -s /workspace/vllm/tests/tpu/test_custom_dispatcher.py \

5
tests/v1/tpu/test_basic.py
View File

@ -15,9 +15,10 @@ if TYPE_CHECKING:
from tests.conftest import VllmRunner from tests.conftest import VllmRunner
MODELS = [ MODELS = [
"Qwen/Qwen2.5-1.5B-Instruct",
# TODO: Enable this models with v6e
# "Qwen/Qwen2-7B-Instruct", # "Qwen/Qwen2-7B-Instruct",
"meta-llama/Llama-3.1-8B", # "meta-llama/Llama-3.1-8B",
# TODO: Add models here as necessary
] ]
TENSOR_PARALLEL_SIZES = [1] TENSOR_PARALLEL_SIZES = [1]