| 1 | # SPDX-License-Identifier: Apache-2.0 |
| 2 | # SPDX-FileCopyrightText: Copyright contributors to the vLLM project |
| 3 | |
| 4 | |
| 5 | import pytest |
| 6 | import requests |
| 7 | |
| 8 | from tests.utils import RemoteOpenAIServer |
| 9 | from vllm.entrypoints.pooling.pooling.protocol import PoolingResponse |
| 10 | |
| 11 | MODEL_NAME = "intfloat/multilingual-e5-small" |
| 12 | DTYPE = "bfloat16" |
| 13 | input_text = "The best thing about vLLM is that it supports many different models" |
| 14 | input_tokens = [ |
| 15 | 0, |
| 16 | 581, |
| 17 | 2965, |
| 18 | 13580, |
| 19 | 1672, |
| 20 | 81, |
| 21 | 23708, |
| 22 | 594, |
| 23 | 83, |
| 24 | 450, |
| 25 | 442, |
| 26 | 8060, |
| 27 | 7, |
| 28 | 5941, |
| 29 | 12921, |
| 30 | 115774, |
| 31 | 2, |
| 32 | ] |
| 33 | |
| 34 | |
| 35 | @pytest.fixture(scope="module") |
| 36 | def server(): |
| 37 | args = [ |
| 38 | "--runner", |
| 39 | "pooling", |
| 40 | "--dtype", |
| 41 | DTYPE, |
| 42 | "--enforce-eager", |
| 43 | "--max-model-len", |
| 44 | "512", |
| 45 | "--pooler-config.task", |
| 46 | "token_embed", |
| 47 | ] |
| 48 | |
| 49 | with RemoteOpenAIServer(MODEL_NAME, args) as remote_server: |
| 50 | yield remote_server |
| 51 | |
| 52 | |
| 53 | @pytest.mark.asyncio |
| 54 | @pytest.mark.parametrize("model_name", [MODEL_NAME]) |
| 55 | async def test_pooling_token_embed(server: RemoteOpenAIServer, model_name: str): |
| 56 | task = "token_embed" |
| 57 | request_id = "client-request-id" |
| 58 | response = requests.post( |
| 59 | server.url_for("pooling"), |
| 60 | json={ |
| 61 | "model": model_name, |
| 62 | "input": input_text, |
| 63 | "encoding_format": "float", |
| 64 | "task": task, |
| 65 | "request_id": request_id, |
| 66 | }, |
| 67 | ) |
| 68 | |
| 69 | poolings = PoolingResponse.model_validate(response.json()) |
| 70 | |
| 71 | assert poolings.id == f"pooling-{request_id}" |
| 72 | assert len(poolings.data) == 1 |
| 73 | assert len(poolings.data[0].data) == len(input_tokens) |
| 74 | assert isinstance(poolings.data[0].data[0], list) |
| 75 | assert len(poolings.data[0].data[0]) == 384 |
| 76 | |
| 77 | |
| 78 | @pytest.mark.asyncio |
| 79 | @pytest.mark.parametrize("model_name", [MODEL_NAME]) |
| 80 | @pytest.mark.parametrize("task", ["embed", "classify", "token_classify", "plugin"]) |
| 81 | async def test_pooling_not_supported( |
| 82 | server: RemoteOpenAIServer, model_name: str, task: str |
| 83 | ): |
| 84 | response = requests.post( |
| 85 | server.url_for("pooling"), |
| 86 | json={ |
| 87 | "model": model_name, |
| 88 | "input": "test", |
| 89 | "encoding_format": "float", |
| 90 | "task": task, |
| 91 | }, |
| 92 | ) |
| 93 | assert response.json()["error"]["type"] == "BadRequestError" |
| 94 | |
| 95 | if task == "plugin": |
| 96 | err_msg = "No IOProcessor plugin installed." |
| 97 | elif task == "embed": |
| 98 | err_msg = "Try switching the model's pooling_task via" |
| 99 | else: |
| 100 | err_msg = f"Unsupported task: {task!r}" |
| 101 | assert response.json()["error"]["message"].startswith(err_msg) |