Skip to content

vllm.entrypoints.pooling.pooling.protocol

PoolingRequest module-attribute 

PoolingRequest: TypeAlias = (
    PoolingCompletionRequest
    | PoolingChatRequest
    | IOProcessorRequest
)

T module-attribute 

T = TypeVar('T')

logger module-attribute 

logger = init_logger(__name__)

IOProcessorRequest

Bases: PoolingBasicRequestMixin, EncodingRequestMixin, Generic[T]

Source code in vllm/entrypoints/pooling/pooling/protocol.py 
102
103
104
105
106
107
class IOProcessorRequest(PoolingBasicRequestMixin, EncodingRequestMixin, Generic[T]):
    data: T
    task: PoolingTask = "plugin"

    def to_pooling_params(self):
        return PoolingParams()

data instance-attribute 

data: T

task class-attribute instance-attribute 

task: PoolingTask = 'plugin'

to_pooling_params 

to_pooling_params()
Source code in vllm/entrypoints/pooling/pooling/protocol.py 
106
107
def to_pooling_params(self):
    return PoolingParams()

IOProcessorResponse

Bases: OpenAIBaseModel, Generic[T]

Source code in vllm/entrypoints/pooling/pooling/protocol.py 
110
111
112
113
114
115
116
117
118
119
120
121
class IOProcessorResponse(OpenAIBaseModel, Generic[T]):
    request_id: str | None = None
    """
    The request_id associated with this response
    """
    created_at: int = Field(default_factory=lambda: int(time.time()))

    data: T
    """
    When using plugins IOProcessor plugins, the actual output is generated
    by the plugin itself. Hence, we use a generic type for the response data
    """

created_at class-attribute instance-attribute 

created_at: int = Field(default_factory=lambda: int(time()))

data instance-attribute 

data: T

When using plugins IOProcessor plugins, the actual output is generated by the plugin itself. Hence, we use a generic type for the response data

request_id class-attribute instance-attribute 

request_id: str | None = None

The request_id associated with this response

PoolingBytesResponse

Bases: OpenAIBaseModel

Source code in vllm/entrypoints/pooling/pooling/protocol.py 
144
145
146
147
class PoolingBytesResponse(OpenAIBaseModel):
    content: list[bytes]
    headers: dict[str, str] | None = None
    media_type: str = "application/octet-stream"

content instance-attribute 

content: list[bytes]

headers class-attribute instance-attribute 

headers: dict[str, str] | None = None

media_type class-attribute instance-attribute 

media_type: str = 'application/octet-stream'

PoolingChatRequest

Bases: PoolingBasicRequestMixin, ChatRequestMixin, EmbedRequestMixin, ClassifyRequestMixin

Source code in vllm/entrypoints/pooling/pooling/protocol.py 
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
class PoolingChatRequest(
    PoolingBasicRequestMixin, ChatRequestMixin, EmbedRequestMixin, ClassifyRequestMixin
):
    task: PoolingTask | None = None

    mm_processor_kwargs: dict[str, Any] | None = Field(
        default=None,
        description=("Additional kwargs to pass to the HF processor."),
    )

    def build_tok_params(self, model_config: ModelConfig) -> TokenizeParams:
        encoder_config = model_config.encoder_config or {}

        return TokenizeParams(
            max_total_tokens=model_config.max_model_len,
            max_output_tokens=0,
            truncate_prompt_tokens=self.truncate_prompt_tokens,
            do_lower_case=encoder_config.get("do_lower_case", False),
            add_special_tokens=self.add_special_tokens,
            max_total_tokens_param="max_model_len",
        )

    def to_pooling_params(self):
        if self.normalize is not None:
            logger.warning_once(
                "`normalize` is deprecated and will be removed in v0.17. "
                "Please pass `use_activation` instead."
            )
            self.use_activation = self.normalize

        return PoolingParams(
            truncate_prompt_tokens=self.truncate_prompt_tokens,
            use_activation=self.use_activation,
            dimensions=self.dimensions,
        )

mm_processor_kwargs class-attribute instance-attribute 

mm_processor_kwargs: dict[str, Any] | None = Field(
    default=None,
    description="Additional kwargs to pass to the HF processor.",
)

task class-attribute instance-attribute 

task: PoolingTask | None = None

build_tok_params 

build_tok_params(
    model_config: ModelConfig,
) -> TokenizeParams
Source code in vllm/entrypoints/pooling/pooling/protocol.py 
72
73
74
75
76
77
78
79
80
81
82
def build_tok_params(self, model_config: ModelConfig) -> TokenizeParams:
    encoder_config = model_config.encoder_config or {}

    return TokenizeParams(
        max_total_tokens=model_config.max_model_len,
        max_output_tokens=0,
        truncate_prompt_tokens=self.truncate_prompt_tokens,
        do_lower_case=encoder_config.get("do_lower_case", False),
        add_special_tokens=self.add_special_tokens,
        max_total_tokens_param="max_model_len",
    )

to_pooling_params 

to_pooling_params()
Source code in vllm/entrypoints/pooling/pooling/protocol.py 
84
85
86
87
88
89
90
91
92
93
94
95
96
def to_pooling_params(self):
    if self.normalize is not None:
        logger.warning_once(
            "`normalize` is deprecated and will be removed in v0.17. "
            "Please pass `use_activation` instead."
        )
        self.use_activation = self.normalize

    return PoolingParams(
        truncate_prompt_tokens=self.truncate_prompt_tokens,
        use_activation=self.use_activation,
        dimensions=self.dimensions,
    )

PoolingCompletionRequest

Bases: PoolingBasicRequestMixin, CompletionRequestMixin, EmbedRequestMixin, ClassifyRequestMixin

Source code in vllm/entrypoints/pooling/pooling/protocol.py 
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
class PoolingCompletionRequest(
    PoolingBasicRequestMixin,
    CompletionRequestMixin,
    EmbedRequestMixin,
    ClassifyRequestMixin,
):
    task: PoolingTask | None = None

    def build_tok_params(self, model_config: ModelConfig) -> TokenizeParams:
        encoder_config = model_config.encoder_config or {}

        return TokenizeParams(
            max_total_tokens=model_config.max_model_len,
            max_output_tokens=0,
            truncate_prompt_tokens=self.truncate_prompt_tokens,
            do_lower_case=encoder_config.get("do_lower_case", False),
            add_special_tokens=self.add_special_tokens,
            max_total_tokens_param="max_model_len",
        )

    def to_pooling_params(self):
        if self.normalize is not None:
            logger.warning_once(
                "`normalize` is deprecated and will be removed in v0.17. "
                "Please pass `use_activation` instead."
            )
            self.use_activation = self.normalize

        return PoolingParams(
            truncate_prompt_tokens=self.truncate_prompt_tokens,
            use_activation=self.use_activation,
            dimensions=self.dimensions,
        )

task class-attribute instance-attribute 

task: PoolingTask | None = None

build_tok_params 

build_tok_params(
    model_config: ModelConfig,
) -> TokenizeParams
Source code in vllm/entrypoints/pooling/pooling/protocol.py 
35
36
37
38
39
40
41
42
43
44
45
def build_tok_params(self, model_config: ModelConfig) -> TokenizeParams:
    encoder_config = model_config.encoder_config or {}

    return TokenizeParams(
        max_total_tokens=model_config.max_model_len,
        max_output_tokens=0,
        truncate_prompt_tokens=self.truncate_prompt_tokens,
        do_lower_case=encoder_config.get("do_lower_case", False),
        add_special_tokens=self.add_special_tokens,
        max_total_tokens_param="max_model_len",
    )

to_pooling_params 

to_pooling_params()
Source code in vllm/entrypoints/pooling/pooling/protocol.py 
47
48
49
50
51
52
53
54
55
56
57
58
59
def to_pooling_params(self):
    if self.normalize is not None:
        logger.warning_once(
            "`normalize` is deprecated and will be removed in v0.17. "
            "Please pass `use_activation` instead."
        )
        self.use_activation = self.normalize

    return PoolingParams(
        truncate_prompt_tokens=self.truncate_prompt_tokens,
        use_activation=self.use_activation,
        dimensions=self.dimensions,
    )

PoolingResponse

Bases: OpenAIBaseModel

Source code in vllm/entrypoints/pooling/pooling/protocol.py 
135
136
137
138
139
140
141
class PoolingResponse(OpenAIBaseModel):
    id: str = Field(default_factory=lambda: f"pool-{random_uuid()}")
    object: str = "list"
    created: int = Field(default_factory=lambda: int(time.time()))
    model: str
    data: list[PoolingResponseData]
    usage: UsageInfo

created class-attribute instance-attribute 

created: int = Field(default_factory=lambda: int(time()))

data instance-attribute 

data: list[PoolingResponseData]

id class-attribute instance-attribute 

id: str = Field(
    default_factory=lambda: f"pool-{random_uuid()}"
)

model instance-attribute 

model: str

object class-attribute instance-attribute 

object: str = 'list'

usage instance-attribute 

usage: UsageInfo

PoolingResponseData

Bases: OpenAIBaseModel

Source code in vllm/entrypoints/pooling/pooling/protocol.py 
129
130
131
132
class PoolingResponseData(OpenAIBaseModel):
    index: int
    object: str = "pooling"
    data: list[list[float]] | list[float] | str

data instance-attribute 

data: list[list[float]] | list[float] | str

index instance-attribute 

index: int

object class-attribute instance-attribute 

object: str = 'pooling'