Skip to content

vllm.entrypoints.pooling.score.protocol

ScoreRequest module-attribute 

ScoreRequest: TypeAlias = (
    ScoreQueriesDocumentsRequest
    | ScoreDataRequest
    | ScoreTextRequest
)

RerankDocument

Bases: BaseModel

Source code in vllm/entrypoints/pooling/score/protocol.py 
109
110
111
class RerankDocument(BaseModel):
    text: str | None = None
    multi_modal: ScoreContentPartParam | None = None

multi_modal class-attribute instance-attribute 

multi_modal: ScoreContentPartParam | None = None

text class-attribute instance-attribute 

text: str | None = None

RerankRequest

Bases: PoolingBasicRequestMixin, ClassifyRequestMixin

Source code in vllm/entrypoints/pooling/score/protocol.py 
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
class RerankRequest(PoolingBasicRequestMixin, ClassifyRequestMixin):
    query: str | ScoreMultiModalParam
    documents: list[str] | ScoreMultiModalParam
    top_n: int = Field(default_factory=lambda: 0)

    # --8<-- [start:rerank-extra-params]
    mm_processor_kwargs: dict[str, Any] | None = Field(
        default=None,
        description=("Additional kwargs to pass to the HF processor."),
    )
    # --8<-- [end:rerank-extra-params]

    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),
            max_total_tokens_param="max_model_len",
        )

documents instance-attribute 

documents: list[str] | ScoreMultiModalParam

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.",
)

query instance-attribute 

query: str | ScoreMultiModalParam

top_n class-attribute instance-attribute 

top_n: int = Field(default_factory=lambda: 0)

build_tok_params 

build_tok_params(
    model_config: ModelConfig,
) -> TokenizeParams
Source code in vllm/entrypoints/pooling/score/protocol.py 
 97
 98
 99
100
101
102
103
104
105
106
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),
        max_total_tokens_param="max_model_len",
    )

RerankResponse

Bases: OpenAIBaseModel

Source code in vllm/entrypoints/pooling/score/protocol.py 
125
126
127
128
129
class RerankResponse(OpenAIBaseModel):
    id: str
    model: str
    usage: RerankUsage
    results: list[RerankResult]

id instance-attribute 

id: str

model instance-attribute 

model: str

results instance-attribute 

results: list[RerankResult]

usage instance-attribute 

usage: RerankUsage

RerankResult

Bases: BaseModel

Source code in vllm/entrypoints/pooling/score/protocol.py 
114
115
116
117
class RerankResult(BaseModel):
    index: int
    document: RerankDocument
    relevance_score: float

document instance-attribute 

document: RerankDocument

index instance-attribute 

index: int

relevance_score instance-attribute 

relevance_score: float

RerankUsage

Bases: BaseModel

Source code in vllm/entrypoints/pooling/score/protocol.py 
120
121
122
class RerankUsage(BaseModel):
    prompt_tokens: int
    total_tokens: int

prompt_tokens instance-attribute 

prompt_tokens: int

total_tokens instance-attribute 

total_tokens: int

ScoreDataRequest

Bases: ScoreRequestMixin

Source code in vllm/entrypoints/pooling/score/protocol.py 
49
50
51
class ScoreDataRequest(ScoreRequestMixin):
    data_1: list[str] | str | ScoreMultiModalParam
    data_2: list[str] | str | ScoreMultiModalParam

data_1 instance-attribute 

data_1: list[str] | str | ScoreMultiModalParam

data_2 instance-attribute 

data_2: list[str] | str | ScoreMultiModalParam

ScoreQueriesDocumentsRequest

Bases: ScoreRequestMixin

Source code in vllm/entrypoints/pooling/score/protocol.py 
54
55
56
57
58
59
60
61
62
63
64
class ScoreQueriesDocumentsRequest(ScoreRequestMixin):
    queries: list[str] | str | ScoreMultiModalParam
    documents: list[str] | str | ScoreMultiModalParam

    @property
    def data_1(self):
        return self.queries

    @property
    def data_2(self):
        return self.documents

data_1 property 

data_1

data_2 property 

data_2

documents instance-attribute 

documents: list[str] | str | ScoreMultiModalParam

queries instance-attribute 

queries: list[str] | str | ScoreMultiModalParam

ScoreRequestMixin

Bases: PoolingBasicRequestMixin, ClassifyRequestMixin

Source code in vllm/entrypoints/pooling/score/protocol.py 
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
class ScoreRequestMixin(PoolingBasicRequestMixin, ClassifyRequestMixin):
    # --8<-- [start:score-extra-params]
    mm_processor_kwargs: dict[str, Any] | None = Field(
        default=None,
        description=("Additional kwargs to pass to the HF processor."),
    )
    # --8<-- [end:score-extra-params]

    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),
            max_total_tokens_param="max_model_len",
        )

    def to_pooling_params(self):
        return PoolingParams(
            truncate_prompt_tokens=self.truncate_prompt_tokens,
            use_activation=self.use_activation,
        )

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.",
)

build_tok_params 

build_tok_params(
    model_config: ModelConfig,
) -> TokenizeParams
Source code in vllm/entrypoints/pooling/score/protocol.py 
31
32
33
34
35
36
37
38
39
40
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),
        max_total_tokens_param="max_model_len",
    )

to_pooling_params 

to_pooling_params()
Source code in vllm/entrypoints/pooling/score/protocol.py 
42
43
44
45
46
def to_pooling_params(self):
    return PoolingParams(
        truncate_prompt_tokens=self.truncate_prompt_tokens,
        use_activation=self.use_activation,
    )

ScoreResponse

Bases: OpenAIBaseModel

Source code in vllm/entrypoints/pooling/score/protocol.py 
138
139
140
141
142
143
144
class ScoreResponse(OpenAIBaseModel):
    id: str = Field(default_factory=lambda: f"embd-{random_uuid()}")
    object: str = "list"
    created: int = Field(default_factory=lambda: int(time.time()))
    model: str
    data: list[ScoreResponseData]
    usage: UsageInfo

created class-attribute instance-attribute 

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

data instance-attribute 

data: list[ScoreResponseData]

id class-attribute instance-attribute 

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

model instance-attribute 

model: str

object class-attribute instance-attribute 

object: str = 'list'

usage instance-attribute 

usage: UsageInfo

ScoreResponseData

Bases: OpenAIBaseModel

Source code in vllm/entrypoints/pooling/score/protocol.py 
132
133
134
135
class ScoreResponseData(OpenAIBaseModel):
    index: int
    object: str = "score"
    score: float

index instance-attribute 

index: int

object class-attribute instance-attribute 

object: str = 'score'

score instance-attribute 

score: float

ScoreTextRequest

Bases: ScoreRequestMixin

Source code in vllm/entrypoints/pooling/score/protocol.py 
67
68
69
70
71
72
73
74
75
76
77
class ScoreTextRequest(ScoreRequestMixin):
    text_1: list[str] | str | ScoreMultiModalParam
    text_2: list[str] | str | ScoreMultiModalParam

    @property
    def data_1(self):
        return self.text_1

    @property
    def data_2(self):
        return self.text_2

data_1 property 

data_1

data_2 property 

data_2

text_1 instance-attribute 

text_1: list[str] | str | ScoreMultiModalParam

text_2 instance-attribute 

text_2: list[str] | str | ScoreMultiModalParam