xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2026-06-04 20:22:15 +08:00
Code Issues Packages Projects Releases Wiki Activity

Use smaller embedding model when not testing model specifically (#13891)

Browse Source
This commit is contained in:
Harry Mellor 2025-02-28 08:50:43 +00:00 committed by GitHub
parent b9e41734c5
commit 76c89fcadd
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
9 changed files with 15 additions and 15 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
tests/entrypoints/llm/test_encode.py
View File

@ -8,7 +8,7 @@ import pytest
from vllm import LLM, PoolingParams, PoolingRequestOutput from vllm import LLM, PoolingParams, PoolingRequestOutput
from vllm.distributed import cleanup_dist_env_and_memory from vllm.distributed import cleanup_dist_env_and_memory
MODEL_NAME = "intfloat/e5-mistral-7b-instruct" MODEL_NAME = "intfloat/multilingual-e5-small"
PROMPTS = [ PROMPTS = [
"Hello, my name is", "Hello, my name is",

2
tests/entrypoints/openai/test_embedding.py
View File

@ -13,7 +13,7 @@ from vllm.transformers_utils.tokenizer import get_tokenizer
from ...utils import RemoteOpenAIServer from ...utils import RemoteOpenAIServer
MODEL_NAME = "intfloat/e5-mistral-7b-instruct" MODEL_NAME = "intfloat/multilingual-e5-small"
DUMMY_CHAT_TEMPLATE = """{% for message in messages %}{{message['role'] + ': ' + message['content'] + '\\n'}}{% endfor %}""" # noqa: E501 DUMMY_CHAT_TEMPLATE = """{% for message in messages %}{{message['role'] + ': ' + message['content'] + '\\n'}}{% endfor %}""" # noqa: E501

4
tests/entrypoints/openai/test_metrics.py
View File

@ -282,7 +282,7 @@ async def test_metrics_exist(server: RemoteOpenAIServer,
def test_metrics_exist_run_batch(use_v1: bool): def test_metrics_exist_run_batch(use_v1: bool):
if use_v1: if use_v1:
pytest.skip("Skipping test on vllm V1") pytest.skip("Skipping test on vllm V1")
input_batch = """{"custom_id": "request-0", "method": "POST", "url": "/v1/embeddings", "body": {"model": "intfloat/e5-mistral-7b-instruct", "input": "You are a helpful assistant."}}""" # noqa: E501 input_batch = """{"custom_id": "request-0", "method": "POST", "url": "/v1/embeddings", "body": {"model": "intfloat/multilingual-e5-small", "input": "You are a helpful assistant."}}""" # noqa: E501
base_url = "0.0.0.0" base_url = "0.0.0.0"
port = "8001" port = "8001"
@ -302,7 +302,7 @@ def test_metrics_exist_run_batch(use_v1: bool):
"-o", "-o",
output_file.name, output_file.name,
"--model", "--model",
"intfloat/e5-mistral-7b-instruct", "intfloat/multilingual-e5-small",
"--enable-metrics", "--enable-metrics",
"--url", "--url",
base_url, base_url,

10
tests/entrypoints/openai/test_run_batch.py
View File

@ -18,10 +18,10 @@ INPUT_BATCH = """{"custom_id": "request-1", "method": "POST", "url": "/v1/chat/c
INVALID_INPUT_BATCH = """{"invalid_field": "request-1", "method": "POST", "url": "/v1/chat/completions", "body": {"model": "NousResearch/Meta-Llama-3-8B-Instruct", "messages": [{"role": "system", "content": "You are a helpful assistant."},{"role": "user", "content": "Hello world!"}],"max_tokens": 1000}} INVALID_INPUT_BATCH = """{"invalid_field": "request-1", "method": "POST", "url": "/v1/chat/completions", "body": {"model": "NousResearch/Meta-Llama-3-8B-Instruct", "messages": [{"role": "system", "content": "You are a helpful assistant."},{"role": "user", "content": "Hello world!"}],"max_tokens": 1000}}
{"custom_id": "request-2", "method": "POST", "url": "/v1/chat/completions", "body": {"model": "NousResearch/Meta-Llama-3-8B-Instruct", "messages": [{"role": "system", "content": "You are an unhelpful assistant."},{"role": "user", "content": "Hello world!"}],"max_tokens": 1000}}""" {"custom_id": "request-2", "method": "POST", "url": "/v1/chat/completions", "body": {"model": "NousResearch/Meta-Llama-3-8B-Instruct", "messages": [{"role": "system", "content": "You are an unhelpful assistant."},{"role": "user", "content": "Hello world!"}],"max_tokens": 1000}}"""
INPUT_EMBEDDING_BATCH = """{"custom_id": "request-1", "method": "POST", "url": "/v1/embeddings", "body": {"model": "intfloat/e5-mistral-7b-instruct", "input": "You are a helpful assistant."}} INPUT_EMBEDDING_BATCH = """{"custom_id": "request-1", "method": "POST", "url": "/v1/embeddings", "body": {"model": "intfloat/multilingual-e5-small", "input": "You are a helpful assistant."}}
{"custom_id": "request-2", "method": "POST", "url": "/v1/embeddings", "body": {"model": "intfloat/e5-mistral-7b-instruct", "input": "You are an unhelpful assistant."}} {"custom_id": "request-2", "method": "POST", "url": "/v1/embeddings", "body": {"model": "intfloat/multilingual-e5-small", "input": "You are an unhelpful assistant."}}
{"custom_id": "request-3", "method": "POST", "url": "/v1/embeddings", "body": {"model": "intfloat/e5-mistral-7b-instruct", "input": "Hello world!"}} {"custom_id": "request-3", "method": "POST", "url": "/v1/embeddings", "body": {"model": "intfloat/multilingual-e5-small", "input": "Hello world!"}}
{"custom_id": "request-4", "method": "POST", "url": "/v1/embeddings", "body": {"model": "NonExistModel", "input": "Hello world!"}}""" {"custom_id": "request-4", "method": "POST", "url": "/v1/embeddings", "body": {"model": "NonExistModel", "input": "Hello world!"}}"""
INPUT_SCORE_BATCH = """{"custom_id": "request-1", "method": "POST", "url": "/v1/score", "body": {"model": "BAAI/bge-reranker-v2-m3", "text_1": "What is the capital of France?", "text_2": ["The capital of Brazil is Brasilia.", "The capital of France is Paris."]}} INPUT_SCORE_BATCH = """{"custom_id": "request-1", "method": "POST", "url": "/v1/score", "body": {"model": "BAAI/bge-reranker-v2-m3", "text_1": "What is the capital of France?", "text_2": ["The capital of Brazil is Brasilia.", "The capital of France is Paris."]}}
@ -37,7 +37,7 @@ def test_empty_file():
proc = subprocess.Popen([ proc = subprocess.Popen([
sys.executable, "-m", "vllm.entrypoints.openai.run_batch", "-i", sys.executable, "-m", "vllm.entrypoints.openai.run_batch", "-i",
input_file.name, "-o", output_file.name, "--model", input_file.name, "-o", output_file.name, "--model",
"intfloat/e5-mistral-7b-instruct" "intfloat/multilingual-e5-small"
], ) ], )
proc.communicate() proc.communicate()
proc.wait() proc.wait()
@ -97,7 +97,7 @@ def test_embeddings():
proc = subprocess.Popen([ proc = subprocess.Popen([
sys.executable, "-m", "vllm.entrypoints.openai.run_batch", "-i", sys.executable, "-m", "vllm.entrypoints.openai.run_batch", "-i",
input_file.name, "-o", output_file.name, "--model", input_file.name, "-o", output_file.name, "--model",
"intfloat/e5-mistral-7b-instruct" "intfloat/multilingual-e5-small"
], ) ], )
proc.communicate() proc.communicate()
proc.wait() proc.wait()

4
tests/model_executor/test_model_load_with_params.py
View File

@ -14,7 +14,7 @@ MODEL_NAME = os.environ.get("MODEL_NAME", "BAAI/bge-base-en-v1.5")
REVISION = os.environ.get("REVISION", "main") REVISION = os.environ.get("REVISION", "main")
MODEL_NAME_ROBERTA = os.environ.get("MODEL_NAME", MODEL_NAME_ROBERTA = os.environ.get("MODEL_NAME",
"intfloat/multilingual-e5-large") "intfloat/multilingual-e5-small")
REVISION_ROBERTA = os.environ.get("REVISION", "main") REVISION_ROBERTA = os.environ.get("REVISION", "main")
@ -83,7 +83,7 @@ def test_roberta_model_loading_with_params(vllm_runner):
assert model_config.pooler_config.pooling_norm assert model_config.pooler_config.pooling_norm
# asserts on the tokenizer loaded # asserts on the tokenizer loaded
assert model_tokenizer.tokenizer_id == "intfloat/multilingual-e5-large" assert model_tokenizer.tokenizer_id == "intfloat/multilingual-e5-small"
assert not model_tokenizer.tokenizer_config["do_lower_case"] assert not model_tokenizer.tokenizer_config["do_lower_case"]
def check_model(model): def check_model(model):

2
tests/models/embedding/language/test_embedding.py
View File

@ -17,7 +17,7 @@ from ..utils import check_embeddings_close
pytest.param("BAAI/bge-base-en-v1.5", pytest.param("BAAI/bge-base-en-v1.5",
marks=[pytest.mark.core_model, pytest.mark.cpu_model]), marks=[pytest.mark.core_model, pytest.mark.cpu_model]),
pytest.param("sentence-transformers/all-MiniLM-L12-v2"), pytest.param("sentence-transformers/all-MiniLM-L12-v2"),
pytest.param("intfloat/multilingual-e5-large"), pytest.param("intfloat/multilingual-e5-small"),
# [Decoder-only] # [Decoder-only]
pytest.param("BAAI/bge-multilingual-gemma2", pytest.param("BAAI/bge-multilingual-gemma2",
marks=[pytest.mark.core_model]), marks=[pytest.mark.core_model]),

2
tests/models/registry.py
View File

@ -211,7 +211,7 @@ _EMBEDDING_EXAMPLE_MODELS = {
"Qwen2ForSequenceClassification": _HfExamplesInfo("jason9693/Qwen2.5-1.5B-apeach"), # noqa: E501 "Qwen2ForSequenceClassification": _HfExamplesInfo("jason9693/Qwen2.5-1.5B-apeach"), # noqa: E501
"RobertaModel": _HfExamplesInfo("sentence-transformers/stsb-roberta-base-v2"), # noqa: E501 "RobertaModel": _HfExamplesInfo("sentence-transformers/stsb-roberta-base-v2"), # noqa: E501
"RobertaForMaskedLM": _HfExamplesInfo("sentence-transformers/all-roberta-large-v1"), # noqa: E501 "RobertaForMaskedLM": _HfExamplesInfo("sentence-transformers/all-roberta-large-v1"), # noqa: E501
"XLMRobertaModel": _HfExamplesInfo("intfloat/multilingual-e5-large"), "XLMRobertaModel": _HfExamplesInfo("intfloat/multilingual-e5-small"),
# [Multimodal] # [Multimodal]
"LlavaNextForConditionalGeneration": _HfExamplesInfo("royokong/e5-v"), "LlavaNextForConditionalGeneration": _HfExamplesInfo("royokong/e5-v"),
"Phi3VForCausalLM": _HfExamplesInfo("TIGER-Lab/VLM2Vec-Full", "Phi3VForCausalLM": _HfExamplesInfo("TIGER-Lab/VLM2Vec-Full",

2
tests/test_config.py
View File

@ -13,7 +13,7 @@ from vllm.platforms import current_platform
("model_id", "expected_runner_type", "expected_task"), ("model_id", "expected_runner_type", "expected_task"),
[ [
("distilbert/distilgpt2", "generate", "generate"), ("distilbert/distilgpt2", "generate", "generate"),
("intfloat/e5-mistral-7b-instruct", "pooling", "embed"), ("intfloat/multilingual-e5-small", "pooling", "embed"),
("jason9693/Qwen2.5-1.5B-apeach", "pooling", "classify"), ("jason9693/Qwen2.5-1.5B-apeach", "pooling", "classify"),
("cross-encoder/ms-marco-MiniLM-L-6-v2", "pooling", "score"), ("cross-encoder/ms-marco-MiniLM-L-6-v2", "pooling", "score"),
("Qwen/Qwen2.5-Math-RM-72B", "pooling", "reward"), ("Qwen/Qwen2.5-Math-RM-72B", "pooling", "reward"),

2
vllm/test_utils.py
View File

@ -28,7 +28,7 @@ MODELS_ON_S3 = [
"HuggingFaceM4/Idefics3-8B-Llama3", "HuggingFaceM4/Idefics3-8B-Llama3",
"internlm/internlm2-1_8b-reward", "internlm/internlm2-1_8b-reward",
"intfloat/e5-mistral-7b-instruct", "intfloat/e5-mistral-7b-instruct",
"intfloat/multilingual-e5-large", "intfloat/multilingual-e5-small",
"jason9693/Qwen2.5-1.5B-apeach", "jason9693/Qwen2.5-1.5B-apeach",
"llava-hf/llava-1.5-7b-hf", "llava-hf/llava-1.5-7b-hf",
"llava-hf/llava-onevision-qwen2-0.5b-ov-hf", "llava-hf/llava-onevision-qwen2-0.5b-ov-hf",