wassname/vllm
1
0
Fork 0
mirror of https://github.com/wassname/vllm.git synced 2026-06-27 18:27:02 +08:00
Code Issues Packages Projects Releases Wiki Activity

[2/n][ci] S3: Use full model path (#13564)

Browse Source 
Signed-off-by: <>
This commit is contained in:
Kevin H. Luu
2025-02-20 01:20:15 -08:00
committed by GitHub
parent aa1e62d0db
commit a64a84433d
8 changed files with 26 additions and 19 deletions
Download Patch File Download Diff File Expand all files Collapse all files
tests/engine/test_computed_prefix_blocks.py
+2 -1
View File
@@ -10,7 +10,8 @@ from vllm.sampling_params import SamplingParams
from ..conftest import MODEL_WEIGHTS_S3_BUCKET
@pytest.mark.parametrize("model", [f"{MODEL_WEIGHTS_S3_BUCKET}/distilgpt2"])
@pytest.mark.parametrize("model",
[f"{MODEL_WEIGHTS_S3_BUCKET}/distilbert/distilgpt2"])
@pytest.mark.parametrize("block_size", [16])
def test_computed_prefix_blocks(model: str, block_size: int):
# This test checks if we are able to run the engine to completion
tests/engine/test_detokenization.py
+2 -1
View File
@@ -9,7 +9,8 @@ from vllm.sampling_params import SamplingParams
from ..conftest import MODEL_WEIGHTS_S3_BUCKET
@pytest.mark.parametrize("model", [f"{MODEL_WEIGHTS_S3_BUCKET}/distilgpt2"])
@pytest.mark.parametrize("model",
[f"{MODEL_WEIGHTS_S3_BUCKET}/distilbert/distilgpt2"])
def test_computed_prefix_blocks(model: str):
# This test checks if the engine generates completions both with and
# without optional detokenization, that detokenization includes text
tests/engine/test_executor.py
+8 -4
View File
@@ -38,7 +38,8 @@ class CustomUniExecutor(UniProcExecutor):
CustomUniExecutorAsync = CustomUniExecutor
@pytest.mark.parametrize("model", [f"{MODEL_WEIGHTS_S3_BUCKET}/distilgpt2"])
@pytest.mark.parametrize("model",
[f"{MODEL_WEIGHTS_S3_BUCKET}/distilbert/distilgpt2"])
def test_custom_executor_type_checking(model):
with pytest.raises(ValueError):
engine_args = EngineArgs(model=model,
@@ -51,7 +52,8 @@ def test_custom_executor_type_checking(model):
AsyncLLMEngine.from_engine_args(engine_args)
@pytest.mark.parametrize("model", [f"{MODEL_WEIGHTS_S3_BUCKET}/distilgpt2"])
@pytest.mark.parametrize("model",
[f"{MODEL_WEIGHTS_S3_BUCKET}/distilbert/distilgpt2"])
def test_custom_executor(model, tmp_path):
cwd = os.path.abspath(".")
os.chdir(tmp_path)
@@ -75,7 +77,8 @@ def test_custom_executor(model, tmp_path):
os.chdir(cwd)
@pytest.mark.parametrize("model", [f"{MODEL_WEIGHTS_S3_BUCKET}/distilgpt2"])
@pytest.mark.parametrize("model",
[f"{MODEL_WEIGHTS_S3_BUCKET}/distilbert/distilgpt2"])
def test_custom_executor_async(model, tmp_path):
cwd = os.path.abspath(".")
os.chdir(tmp_path)
@@ -103,7 +106,8 @@ def test_custom_executor_async(model, tmp_path):
os.chdir(cwd)
@pytest.mark.parametrize("model", [f"{MODEL_WEIGHTS_S3_BUCKET}/distilgpt2"])
@pytest.mark.parametrize("model",
[f"{MODEL_WEIGHTS_S3_BUCKET}/distilbert/distilgpt2"])
def test_respect_ray(model):
# even for TP=1 and PP=1,
# if users specify ray, we should use ray.
tests/engine/test_skip_tokenizer_init.py
+2 -1
View File
@@ -9,7 +9,8 @@ from vllm.sampling_params import SamplingParams
from ..conftest import MODEL_WEIGHTS_S3_BUCKET
@pytest.mark.parametrize("model", [f"{MODEL_WEIGHTS_S3_BUCKET}/distilgpt2"])
@pytest.mark.parametrize("model",
[f"{MODEL_WEIGHTS_S3_BUCKET}/distilbert/distilgpt2"])
def test_skip_tokenizer_initialization(model: str):
# This test checks if the flag skip_tokenizer_init skips the initialization
# of tokenizer and detokenizer. The generated output is expected to contain