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 6, 09:30 UTC
[Refactor] Relocate entrypoint tests to match serving code structure (#37593)
Signed-off-by: sfeng33 <4florafeng@gmail.com>
Flora Feng committed
Mar 20, 2026 at 01:31 UTC
e2d1c8b5e897e3f367af8bf026d6547f4b602be4
11 files changed
+8
-5
.buildkite/scripts/hardware_ci/run-amd-test.sh
+5
-1
@@ -337,12 +337,16 @@ apply_rocm_test_overrides() {
337
--ignore=entrypoints/openai/completion/test_shutdown.py \
338
--ignore=entrypoints/openai/test_completion.py \
339
--ignore=entrypoints/openai/models/test_models.py \
340
- --ignore=entrypoints/openai/test_lora_adapters.py \
340
--ignore=entrypoints/openai/test_return_tokens_as_ids.py \
341
--ignore=entrypoints/openai/chat_completion/test_root_path.py \
342
--ignore=entrypoints/openai/completion/test_prompt_validation.py "}
343
fi
344
345
+ if [[ $cmds == *" entrypoints/serve"* ]]; then
346
+ cmds="${cmds} \
347
+ --ignore=entrypoints/serve/lora/test_lora_adapters.py"
348
+ fi
349
+
350
if [[ $cmds == *" entrypoints/llm "* ]]; then
351
cmds=${cmds//" entrypoints/llm "/" entrypoints/llm \
352
--ignore=entrypoints/llm/test_chat.py \
tests/entrypoints/openai/responses/test_protocol.py
renamed
tests/entrypoints/openai/responses/test_responses_utils.py
renamed
tests/entrypoints/serve/disagg/__init__.py
tests/entrypoints/serve/disagg/test_serving_tokens.py
renamed
+1
-2
@@ -8,12 +8,11 @@ import pytest
8
import pytest_asyncio
9
from transformers import AutoTokenizer
10
11
+from tests.utils import RemoteOpenAIServer
12
from vllm.config import ModelConfig
13
from vllm.config.utils import getattr_iter
14
from vllm.v1.engine.detokenizer import check_stop_strings
15
15
-from ...utils import RemoteOpenAIServer
16
-
16
MODEL_NAME = "Qwen/Qwen3-0.6B"
17
GEN_ENDPOINT = "/inference/v1/generate"
18
tests/entrypoints/serve/lora/__init__.py
tests/entrypoints/serve/lora/test_lora_adapters.py
renamed
+1
-1
@@ -10,7 +10,7 @@ import openai # use the official client for correctness check
10
import pytest
11
import pytest_asyncio
12
13
-from ...utils import RemoteOpenAIServer
13
+from tests.utils import RemoteOpenAIServer
14
15
# any model with a chat template should work here
16
MODEL_NAME = "Qwen/Qwen3-0.6B"
tests/entrypoints/serve/lora/test_serving_models.py
renamed
tests/entrypoints/serve/render/__init__.py
tests/entrypoints/serve/render/test_launch_render.py
renamed
+1
-1
@@ -6,7 +6,7 @@ import httpx
6
import pytest
7
import pytest_asyncio
8
9
-from ...utils import RemoteLaunchRenderServer
9
+from tests.utils import RemoteLaunchRenderServer
10
11
MODEL_NAME = "hmellor/tiny-random-LlamaForCausalLM"
12