[Frontend][last/5] Make pooling entrypoints request schema consensus. (#31127)

Signed-off-by: wang.yuqi <yuqi.wang@daocloud.io>
This commit is contained in:
wang.yuqi
2026-02-09 06:42:38 +00:00
committed by GitHub
parent 7c233dbb36
commit 22b64948f6
24 changed files with 658 additions and 612 deletions
@@ -17,6 +17,8 @@ from openai.types.chat import ChatCompletionMessageParam
from openai.types.create_embedding_response import CreateEmbeddingResponse
from PIL import Image
from vllm.utils.print_utils import print_embeddings
# Modify OpenAI's API key and API base to use vLLM's API server.
openai_api_key = "EMPTY"
openai_api_base = "http://localhost:8000/v1"
@@ -51,11 +53,6 @@ def create_chat_embeddings(
)
def print_embeddings(embeds):
embeds_trimmed = (str(embeds[:4])[:-1] + ", ...]") if len(embeds) > 4 else embeds
print(f"Embeddings: {embeds_trimmed} (size={len(embeds)})")
def run_clip(client: OpenAI, model: str):
"""
Start the server using:
@@ -105,7 +102,7 @@ def run_dse_qwen2_vl(client: OpenAI, model: str):
--runner pooling \
--trust-remote-code \
--max-model-len 8192 \
--chat-template examples/template_dse_qwen2_vl.jinja
--chat-template examples/pooling/embed/template/dse_qwen2_vl.jinja
"""
response = create_chat_embeddings(
client,
@@ -316,7 +313,7 @@ def run_vlm2vec(client: OpenAI, model: str):
--runner pooling \
--trust-remote-code \
--max-model-len 4096 \
--chat-template examples/template_vlm2vec_phi3v.jinja
--chat-template examples/pooling/embed/template/vlm2vec_phi3v.jinja
"""
response = create_chat_embeddings(