Hugging Face's logo

Hub Python Library documentation

推理类型

Hugging Face's logo
Join the Hugging Face community

and get access to the augmented documentation experience

to get started

推理类型

此页面列出了 Hugging Face Hub 上每个受支持任务可用的类型（例如 dataclasses）。 每个任务均使用 JSON schema 进行定义，类型由这些 schema 生成——并根据 Python 的要求进行了一些定制。 请访问 @huggingface.js/tasks 以查找每个任务的 JSON schema。

库的这一部分仍在开发中，将在未来的版本中不断改进。

audio_classification

类 huggingface_hub.AudioClassificationInput

< >

( inputs: strparameters: huggingface_hub.inference._generated.types.audio_classification.AudioClassificationParameters | None = None )

音频分类推理的输入

类 huggingface_hub.AudioClassificationOutputElement

< >

( label: strscore: float )

音频分类推理的输出

类 huggingface_hub.AudioClassificationParameters

< >

( function_to_apply: typing.Optional[ForwardRef('AudioClassificationOutputTransform')] = Nonetop_k: int | None = None )

用于音频分类的额外推理参数

audio_to_audio

类 huggingface_hub.AudioToAudioInput

< >

( inputs: typing.Any )

用于音频到音频推理的输入

类 huggingface_hub.AudioToAudioOutputElement

< >

( blob: typing.Anycontent_type: strlabel: str )

Audio To Audio 任务的推理输出 带有其标签的生成音频文件。

automatic_speech_recognition

类 huggingface_hub.AutomaticSpeechRecognitionGenerationParameters

< >

( do_sample: bool | None = Noneearly_stopping: typing.Union[bool, ForwardRef('AutomaticSpeechRecognitionEarlyStoppingEnum'), NoneType] = Noneepsilon_cutoff: float | None = Noneeta_cutoff: float | None = Nonemax_length: int | None = Nonemax_new_tokens: int | None = Nonemin_length: int | None = Nonemin_new_tokens: int | None = Nonenum_beam_groups: int | None = Nonenum_beams: int | None = Nonepenalty_alpha: float | None = Nonetemperature: float | None = Nonetop_k: int | None = Nonetop_p: float | None = Nonetypical_p: float | None = Noneuse_cache: bool | None = None )

文本生成过程的参数化

类 huggingface_hub.AutomaticSpeechRecognitionInput

< >

( inputs: strparameters: huggingface_hub.inference._generated.types.automatic_speech_recognition.AutomaticSpeechRecognitionParameters | None = None )

自动语音识别推理的输入

类 huggingface_hub.AutomaticSpeechRecognitionOutput

< >

( text: strchunks: list[huggingface_hub.inference._generated.types.automatic_speech_recognition.AutomaticSpeechRecognitionOutputChunk] | None = None )

自动语音识别任务的推理输出

类 huggingface_hub.AutomaticSpeechRecognitionOutputChunk

< >

( text: strtimestamp: list )

class huggingface_hub.AutomaticSpeechRecognitionParameters

< >

( generation_parameters: huggingface_hub.inference._generated.types.automatic_speech_recognition.AutomaticSpeechRecognitionGenerationParameters | None = Nonereturn_timestamps: bool | None = None )

用于自动语音识别的额外推理参数

chat_completion

类 huggingface_hub.ChatCompletionInput

< >

( messages: listfrequency_penalty: float | None = Nonelogit_bias: list[float] | None = Nonelogprobs: bool | None = Nonemax_tokens: int | None = Nonemodel: str | None = Nonen: int | None = Nonepresence_penalty: float | None = Noneresponse_format: typing.Union[huggingface_hub.inference._generated.types.chat_completion.ChatCompletionInputResponseFormatText, huggingface_hub.inference._generated.types.chat_completion.ChatCompletionInputResponseFormatJSONSchema, huggingface_hub.inference._generated.types.chat_completion.ChatCompletionInputResponseFormatJSONObject, NoneType] = Noneseed: int | None = Nonestop: list[str] | None = Nonestream: bool | None = Nonestream_options: huggingface_hub.inference._generated.types.chat_completion.ChatCompletionInputStreamOptions | None = Nonetemperature: float | None = Nonetool_choice: typing.Union[huggingface_hub.inference._generated.types.chat_completion.ChatCompletionInputToolChoiceClass, ForwardRef('ChatCompletionInputToolChoiceEnum'), NoneType] = Nonetool_prompt: str | None = Nonetools: list[huggingface_hub.inference._generated.types.chat_completion.ChatCompletionInputTool] | None = Nonetop_logprobs: int | None = Nonetop_p: float | None = None )

Chat Completion 输入。 根据 TGI 规范自动生成。 更多详情，请查看 https://github.com/huggingface/huggingface.js/blob/main/packages/tasks/scripts/inference-tgi-import.ts。

类 huggingface_hub.ChatCompletionInputFunctionDefinition

< >

( name: strparameters: typing.Anydescription: str | None = None )

类 huggingface_hub.ChatCompletionInputFunctionName

< >

( name: str )

类 huggingface_hub.ChatCompletionInputJSONSchema

< >

( name: strdescription: str | None = Noneschema: dict[str, object] | None = Nonestrict: bool | None = None )

类 huggingface_hub.ChatCompletionInputMessage

< >

( role: strcontent: list[huggingface_hub.inference._generated.types.chat_completion.ChatCompletionInputMessageChunk] | str | None = Nonename: str | None = Nonetool_calls: list[huggingface_hub.inference._generated.types.chat_completion.ChatCompletionInputToolCall] | None = None )

类 huggingface_hub.ChatCompletionInputMessageChunk

< >

( type: ChatCompletionInputMessageChunkTypeimage_url: huggingface_hub.inference._generated.types.chat_completion.ChatCompletionInputURL | None = Nonetext: str | None = None )

类 huggingface_hub.ChatCompletionInputResponseFormatJSONObject

< >

( type: typing.Literal['json_object'] )

类 huggingface_hub.ChatCompletionInputResponseFormatJSONSchema

< >

( type: typing.Literal['json_schema']json_schema: ChatCompletionInputJSONSchema )

类 huggingface_hub.ChatCompletionInputResponseFormatText

< >

( type: typing.Literal['text'] )

类 huggingface_hub.ChatCompletionInputStreamOptions

< >

( include_usage: bool | None = None )

类 huggingface_hub.ChatCompletionInputTool

< >

( function: ChatCompletionInputFunctionDefinitiontype: str )

类 huggingface_hub.ChatCompletionInputToolCall

< >

( function: ChatCompletionInputFunctionDefinitionid: strtype: str )

类 huggingface_hub.ChatCompletionInputToolChoiceClass

< >

( function: ChatCompletionInputFunctionName )

类 huggingface_hub.ChatCompletionInputURL

< >

( url: str )

类 huggingface_hub.ChatCompletionOutput

< >

( choices: listcreated: intid: strmodel: strsystem_fingerprint: strusage: ChatCompletionOutputUsage )

Chat Completion 输出。 根据 TGI 规范自动生成。 更多详情，请查看 https://github.com/huggingface/huggingface.js/blob/main/packages/tasks/scripts/inference-tgi-import.ts。

类 huggingface_hub.ChatCompletionOutputComplete

< >

( finish_reason: strindex: intmessage: ChatCompletionOutputMessagelogprobs: huggingface_hub.inference._generated.types.chat_completion.ChatCompletionOutputLogprobs | None = None )

类 huggingface_hub.ChatCompletionOutputFunctionDefinition

< >

( arguments: strname: strdescription: str | None = None )

类 huggingface_hub.ChatCompletionOutputLogprob

< >

( logprob: floattoken: strtop_logprobs: list )

类 huggingface_hub.ChatCompletionOutputLogprobs

< >

( content: list )

类 huggingface_hub.ChatCompletionOutputMessage

< >

( role: strcontent: str | None = Nonereasoning: str | None = Nonetool_call_id: str | None = Nonetool_calls: list[huggingface_hub.inference._generated.types.chat_completion.ChatCompletionOutputToolCall] | None = None )

类 huggingface_hub.ChatCompletionOutputToolCall

< >

( function: ChatCompletionOutputFunctionDefinitionid: strtype: str )

类 huggingface_hub.ChatCompletionOutputTopLogprob

< >

( logprob: floattoken: str )

类 huggingface_hub.ChatCompletionOutputUsage

< >

( completion_tokens: intprompt_tokens: inttotal_tokens: int )

类 huggingface_hub.ChatCompletionStreamOutput

< >

( choices: listcreated: intid: strmodel: strsystem_fingerprint: strusage: huggingface_hub.inference._generated.types.chat_completion.ChatCompletionStreamOutputUsage | None = None )

Chat Completion 流式输出。 根据 TGI 规范自动生成。 更多详情，请查看 https://github.com/huggingface/huggingface.js/blob/main/packages/tasks/scripts/inference-tgi-import.ts。

类 huggingface_hub.ChatCompletionStreamOutputChoice

< >

( delta: ChatCompletionStreamOutputDeltaindex: intfinish_reason: str | None = Nonelogprobs: huggingface_hub.inference._generated.types.chat_completion.ChatCompletionStreamOutputLogprobs | None = None )

类 huggingface_hub.ChatCompletionStreamOutputDelta

< >

( role: strcontent: str | None = Nonereasoning: str | None = Nonetool_call_id: str | None = Nonetool_calls: list[huggingface_hub.inference._generated.types.chat_completion.ChatCompletionStreamOutputDeltaToolCall] | None = None )

类 huggingface_hub.ChatCompletionStreamOutputDeltaToolCall

< >

( function: ChatCompletionStreamOutputFunctionid: strindex: inttype: str )

类 huggingface_hub.ChatCompletionStreamOutputFunction

< >

( arguments: strname: str | None = None )

类 huggingface_hub.ChatCompletionStreamOutputLogprob

< >

( logprob: floattoken: strtop_logprobs: list )

类 huggingface_hub.ChatCompletionStreamOutputLogprobs

< >

( content: list )

类 huggingface_hub.ChatCompletionStreamOutputTopLogprob

< >

( logprob: floattoken: str )

类 huggingface_hub.ChatCompletionStreamOutputUsage

< >

( completion_tokens: intprompt_tokens: inttotal_tokens: int )

depth_estimation

类 huggingface_hub.DepthEstimationInput

< >

( inputs: typing.Anyparameters: dict[str, typing.Any] | None = None )

深度估计推理的输入

类 huggingface_hub.DepthEstimationOutput

< >

( depth: typing.Anypredicted_depth: typing.Any )

深度估计任务的推理输出

document_question_answering

类 huggingface_hub.DocumentQuestionAnsweringInput

< >

( inputs: DocumentQuestionAnsweringInputDataparameters: huggingface_hub.inference._generated.types.document_question_answering.DocumentQuestionAnsweringParameters | None = None )

文档问答推理的输入

类 huggingface_hub.DocumentQuestionAnsweringInputData

< >

( image: typing.Anyquestion: str )

一个（文档，问题）对用于回答

类 huggingface_hub.DocumentQuestionAnsweringOutputElement

< >

( answer: strend: intscore: floatstart: int )

文档问答任务的推理输出

类 huggingface_hub.DocumentQuestionAnsweringParameters

< >

( doc_stride: int | None = Nonehandle_impossible_answer: bool | None = Nonelang: str | None = Nonemax_answer_len: int | None = Nonemax_question_len: int | None = Nonemax_seq_len: int | None = Nonetop_k: int | None = Noneword_boxes: list[list[float] | str] | None = None )

用于文档问答的额外推理参数

feature_extraction

类 huggingface_hub.FeatureExtractionInput

< >

( inputs: list[str] | strnormalize: bool | None = Noneprompt_name: str | None = Nonetruncate: bool | None = Nonetruncation_direction: typing.Optional[ForwardRef('FeatureExtractionInputTruncationDirection')] = None )

特征提取输入。 根据 TEI 规范自动生成。 更多详情，请查看 https://github.com/huggingface/huggingface.js/blob/main/packages/tasks/scripts/inference-tei-import.ts。

fill_mask

类 huggingface_hub.FillMaskInput

< >

( inputs: strparameters: huggingface_hub.inference._generated.types.fill_mask.FillMaskParameters | None = None )

Fill Mask 推理的输入

class huggingface_hub.FillMaskOutputElement

< >

( score: floatsequence: strtoken: inttoken_str: typing.Anyfill_mask_output_token_str: str | None = None )

Fill Mask 任务的推理输出

class huggingface_hub.FillMaskParameters

< >

( targets: list[str] | None = Nonetop_k: int | None = None )

Fill Mask 的额外推理参数

image_classification

类 huggingface_hub.ImageClassificationInput

< >

( inputs: strparameters: huggingface_hub.inference._generated.types.image_classification.ImageClassificationParameters | None = None )

图像分类推理的输入

类 huggingface_hub.ImageClassificationOutputElement

< >

( label: strscore: float )

图像分类任务的推理输出

class huggingface_hub.ImageClassificationParameters

< >

( function_to_apply: typing.Optional[ForwardRef('ImageClassificationOutputTransform')] = Nonetop_k: int | None = None )

图像分类的额外推理参数

image_segmentation

类 huggingface_hub.ImageSegmentationInput

< >

( inputs: strparameters: huggingface_hub.inference._generated.types.image_segmentation.ImageSegmentationParameters | None = None )

图像分割推理的输入

类 huggingface_hub.ImageSegmentationOutputElement

< >

( label: strmask: strscore: float | None = None )

图像分割任务的推理输出 预测的掩码 / 分割

类 huggingface_hub.ImageSegmentationParameters

< >

( mask_threshold: float | None = Noneoverlap_mask_area_threshold: float | None = Nonesubtask: typing.Optional[ForwardRef('ImageSegmentationSubtask')] = Nonethreshold: float | None = None )

图像分割的额外推理参数

image_text_to_image

类 huggingface_hub.ImageTextToImageInput

< >

( inputs: str | None = Noneparameters: huggingface_hub.inference._generated.types.image_text_to_image.ImageTextToImageParameters | None = None )

用于图像文本到图像推理的输入。必须提供 inputs（图像）或 prompt（在参数中），或两者都提供。

类 huggingface_hub.ImageTextToImageOutput

< >

( image: typing.Any )

图像文本到图像任务的推理输出

类 huggingface_hub.ImageTextToImageParameters

< >

( guidance_scale: float | None = Nonenegative_prompt: str | None = Nonenum_inference_steps: int | None = Noneprompt: str | None = Noneseed: int | None = Nonetarget_size: huggingface_hub.inference._generated.types.image_text_to_image.ImageTextToImageTargetSize | None = None )

用于图像文本到图像的额外推理参数

类 huggingface_hub.ImageTextToImageTargetSize

< >

( height: intwidth: int )

输出图像的像素尺寸。此参数仅受部分提供商和特定模型支持。在不支持的情况下将被忽略。

image_text_to_video

类 huggingface_hub.ImageTextToVideoInput

< >

( inputs: str | None = Noneparameters: huggingface_hub.inference._generated.types.image_text_to_video.ImageTextToVideoParameters | None = None )

用于图像文本转视频推理的输入。必须提供 inputs（图像）或 prompt（在参数中），或两者都提供。

类 huggingface_hub.ImageTextToVideoOutput

< >

( video: typing.Any )

图像文本转视频任务的推理输出

类 huggingface_hub.ImageTextToVideoParameters

< >

( guidance_scale: float | None = Nonenegative_prompt: str | None = Nonenum_frames: float | None = Nonenum_inference_steps: int | None = Noneprompt: str | None = Noneseed: int | None = Nonetarget_size: huggingface_hub.inference._generated.types.image_text_to_video.ImageTextToVideoTargetSize | None = None )

用于图像文本转视频的额外推理参数

类 huggingface_hub.ImageTextToVideoTargetSize

< >

( height: intwidth: int )

输出视频帧的像素尺寸。

image_to_image

类 huggingface_hub.ImageToImageInput

< >

( inputs: strparameters: huggingface_hub.inference._generated.types.image_to_image.ImageToImageParameters | None = None )

图像到图像推理的输入

类 huggingface_hub.ImageToImageOutput

< >

( image: typing.Any )

图像到图像任务的推理输出

类 huggingface_hub.ImageToImageParameters

< >

( guidance_scale: float | None = Nonenegative_prompt: str | None = Nonenum_inference_steps: int | None = Noneprompt: str | None = Nonetarget_size: huggingface_hub.inference._generated.types.image_to_image.ImageToImageTargetSize | None = None )

图像到图像的额外推理参数

类 huggingface_hub.ImageToImageTargetSize

< >

( height: intwidth: int )

输出图像的像素尺寸。此参数仅受部分提供商和特定模型支持。在不支持的情况下将被忽略。

image_to_text

类 huggingface_hub.ImageToTextGenerationParameters

< >

( do_sample: bool | None = Noneearly_stopping: typing.Union[bool, ForwardRef('ImageToTextEarlyStoppingEnum'), NoneType] = Noneepsilon_cutoff: float | None = Noneeta_cutoff: float | None = Nonemax_length: int | None = Nonemax_new_tokens: int | None = Nonemin_length: int | None = Nonemin_new_tokens: int | None = Nonenum_beam_groups: int | None = Nonenum_beams: int | None = Nonepenalty_alpha: float | None = Nonetemperature: float | None = Nonetop_k: int | None = Nonetop_p: float | None = Nonetypical_p: float | None = Noneuse_cache: bool | None = None )

文本生成过程的参数化

类 huggingface_hub.ImageToTextInput

< >

( inputs: typing.Anyparameters: huggingface_hub.inference._generated.types.image_to_text.ImageToTextParameters | None = None )

图像到文本推理的输入

类 huggingface_hub.ImageToTextOutput

< >

( generated_text: typing.Anyimage_to_text_output_generated_text: str | None = None )

图像到文本任务的推理输出

class huggingface_hub.ImageToTextParameters

< >

( generation_parameters: huggingface_hub.inference._generated.types.image_to_text.ImageToTextGenerationParameters | None = Nonemax_new_tokens: int | None = None )

用于图像到文本的额外推理参数

image_to_video

类 huggingface_hub.ImageToVideoInput

< >

( inputs: strparameters: huggingface_hub.inference._generated.types.image_to_video.ImageToVideoParameters | None = None )

图像转视频推理的输入

类 huggingface_hub.ImageToVideoOutput

< >

( video: typing.Any )

图像转视频任务的推理输出

类 huggingface_hub.ImageToVideoParameters

< >

( guidance_scale: float | None = Nonenegative_prompt: str | None = Nonenum_frames: float | None = Nonenum_inference_steps: int | None = Noneprompt: str | None = Noneseed: int | None = Nonetarget_size: huggingface_hub.inference._generated.types.image_to_video.ImageToVideoTargetSize | None = None )

用于图像转视频的额外推理参数

类 huggingface_hub.ImageToVideoTargetSize

< >

( height: intwidth: int )

输出视频帧的像素尺寸。

object_detection

类 huggingface_hub.ObjectDetectionBoundingBox

< >

( xmax: intxmin: intymax: intymin: int )

预测的边界框。坐标相对于输入图像的左上角。

类 huggingface_hub.ObjectDetectionInput

< >

( inputs: strparameters: huggingface_hub.inference._generated.types.object_detection.ObjectDetectionParameters | None = None )

目标检测推理的输入

类 huggingface_hub.ObjectDetectionOutputElement

< >

( box: ObjectDetectionBoundingBoxlabel: strscore: float )

目标检测任务的推理输出

类 huggingface_hub.ObjectDetectionParameters

< >

( threshold: float | None = None )

用于目标检测的额外推理参数

question_answering

类 huggingface_hub.QuestionAnsweringInput

< >

( inputs: QuestionAnsweringInputDataparameters: huggingface_hub.inference._generated.types.question_answering.QuestionAnsweringParameters | None = None )

问答推理的输入

类 huggingface_hub.QuestionAnsweringInputData

< >

( context: strquestion: str )

一个（上下文，问题）对用于回答

类 huggingface_hub.QuestionAnsweringOutputElement

< >

( answer: strend: intscore: floatstart: int )

问答任务的推理输出

类 huggingface_hub.QuestionAnsweringParameters

< >

( align_to_words: bool | None = Nonedoc_stride: int | None = Nonehandle_impossible_answer: bool | None = Nonemax_answer_len: int | None = Nonemax_question_len: int | None = Nonemax_seq_len: int | None = Nonetop_k: int | None = None )

用于问答的额外推理参数

sentence_similarity

类 huggingface_hub.SentenceSimilarityInput

< >

( inputs: SentenceSimilarityInputDataparameters: dict[str, typing.Any] | None = None )

句子相似度推理的输入

类 huggingface_hub.SentenceSimilarityInputData

< >

( sentences: listsource_sentence: str )

摘要

类 huggingface_hub.SummarizationInput

< >

( inputs: strparameters: huggingface_hub.inference._generated.types.summarization.SummarizationParameters | None = None )

用于摘要推理的输入

类 huggingface_hub.SummarizationOutput

< >

( summary_text: str )

摘要任务的推理输出

class huggingface_hub.SummarizationParameters

< >

( clean_up_tokenization_spaces: bool | None = Nonegenerate_parameters: dict[str, typing.Any] | None = Nonetruncation: typing.Optional[ForwardRef('SummarizationTruncationStrategy')] = None )

用于摘要的额外推理参数。

table_question_answering

类 huggingface_hub.TableQuestionAnsweringInput

< >

( inputs: TableQuestionAnsweringInputDataparameters: huggingface_hub.inference._generated.types.table_question_answering.TableQuestionAnsweringParameters | None = None )

表格问答推理的输入

类 huggingface_hub.TableQuestionAnsweringInputData

< >

( question: strtable: dict )

一个（表格，问题）对用于回答

类 huggingface_hub.TableQuestionAnsweringOutputElement

< >

( answer: strcells: listcoordinates: listaggregator: str | None = None )

表格问答任务的推理输出

类 huggingface_hub.TableQuestionAnsweringParameters

< >

( padding: typing.Optional[ForwardRef('Padding')] = Nonesequential: bool | None = Nonetruncation: bool | None = None )

用于表格问答的额外推理参数

text2text_generation

类 huggingface_hub.Text2TextGenerationInput

< >

( inputs: strparameters: huggingface_hub.inference._generated.types.text2text_generation.Text2TextGenerationParameters | None = None )

文本到文本生成推理的输入

类 huggingface_hub.Text2TextGenerationOutput

< >

( generated_text: typing.Anytext2_text_generation_output_generated_text: str | None = None )

Text2text Generation 任务的推理输出

类 huggingface_hub.Text2TextGenerationParameters

< >

( clean_up_tokenization_spaces: bool | None = Nonegenerate_parameters: dict[str, typing.Any] | None = Nonetruncation: typing.Optional[ForwardRef('Text2TextGenerationTruncationStrategy')] = None )

用于 Text2text 生成的额外推理参数

text_classification

类 huggingface_hub.TextClassificationInput

< >

( inputs: strparameters: huggingface_hub.inference._generated.types.text_classification.TextClassificationParameters | None = None )

文本分类推理的输入

类 huggingface_hub.TextClassificationOutputElement

< >

( label: strscore: float )

文本分类任务的推理输出

class huggingface_hub.TextClassificationParameters

< >

( function_to_apply: typing.Optional[ForwardRef('TextClassificationOutputTransform')] = Nonetop_k: int | None = None )

用于文本分类的额外推理参数

text_generation

类 huggingface_hub.TextGenerationInput

< >

( inputs: strparameters: huggingface_hub.inference._generated.types.text_generation.TextGenerationInputGenerateParameters | None = Nonestream: bool | None = None )

文本生成输入。 根据 TGI 规范自动生成。 更多详情，请查看 https://github.com/huggingface/huggingface.js/blob/main/packages/tasks/scripts/inference-tgi-import.ts。

类 huggingface_hub.TextGenerationInputGenerateParameters

< >

( adapter_id: str | None = Nonebest_of: int | None = Nonedecoder_input_details: bool | None = Nonedetails: bool | None = Nonedo_sample: bool | None = Nonefrequency_penalty: float | None = Nonegrammar: huggingface_hub.inference._generated.types.text_generation.TextGenerationInputGrammarType | None = Nonemax_new_tokens: int | None = Nonerepetition_penalty: float | None = Nonereturn_full_text: bool | None = Noneseed: int | None = Nonestop: list[str] | None = Nonetemperature: float | None = Nonetop_k: int | None = Nonetop_n_tokens: int | None = Nonetop_p: float | None = Nonetruncate: int | None = Nonetypical_p: float | None = Nonewatermark: bool | None = None )

类 huggingface_hub.TextGenerationInputGrammarType

< >

( type: TypeEnumvalue: typing.Any )

类 huggingface_hub.TextGenerationOutput

< >

( generated_text: strdetails: huggingface_hub.inference._generated.types.text_generation.TextGenerationOutputDetails | None = None )

文本生成输出。 根据 TGI 规范自动生成。 更多详情，请查看 https://github.com/huggingface/huggingface.js/blob/main/packages/tasks/scripts/inference-tgi-import.ts。

类 huggingface_hub.TextGenerationOutputBestOfSequence

< >

( finish_reason: TextGenerationOutputFinishReasongenerated_text: strgenerated_tokens: intprefill: listtokens: listseed: int | None = Nonetop_tokens: list[list[huggingface_hub.inference._generated.types.text_generation.TextGenerationOutputToken]] | None = None )

类 huggingface_hub.TextGenerationOutputDetails

< >

( finish_reason: TextGenerationOutputFinishReasongenerated_tokens: intprefill: listtokens: listbest_of_sequences: list[huggingface_hub.inference._generated.types.text_generation.TextGenerationOutputBestOfSequence] | None = Noneseed: int | None = Nonetop_tokens: list[list[huggingface_hub.inference._generated.types.text_generation.TextGenerationOutputToken]] | None = None )

类 huggingface_hub.TextGenerationOutputPrefillToken

< >

( id: intlogprob: floattext: str )

类 huggingface_hub.TextGenerationOutputToken

< >

( id: intlogprob: floatspecial: booltext: str )

类 huggingface_hub.TextGenerationStreamOutput

< >

( index: inttoken: TextGenerationStreamOutputTokendetails: huggingface_hub.inference._generated.types.text_generation.TextGenerationStreamOutputStreamDetails | None = Nonegenerated_text: str | None = Nonetop_tokens: list[huggingface_hub.inference._generated.types.text_generation.TextGenerationStreamOutputToken] | None = None )

文本生成流式输出。 根据 TGI 规范自动生成。 更多详情，请查看 https://github.com/huggingface/huggingface.js/blob/main/packages/tasks/scripts/inference-tgi-import.ts。

类 huggingface_hub.TextGenerationStreamOutputStreamDetails

< >

( finish_reason: TextGenerationOutputFinishReasongenerated_tokens: intinput_length: intseed: int | None = None )

类 huggingface_hub.TextGenerationStreamOutputToken

< >

( id: intlogprob: floatspecial: booltext: str )

text_to_audio

类 huggingface_hub.TextToAudioGenerationParameters

< >

( do_sample: bool | None = Noneearly_stopping: typing.Union[bool, ForwardRef('TextToAudioEarlyStoppingEnum'), NoneType] = Noneepsilon_cutoff: float | None = Noneeta_cutoff: float | None = Nonemax_length: int | None = Nonemax_new_tokens: int | None = Nonemin_length: int | None = Nonemin_new_tokens: int | None = Nonenum_beam_groups: int | None = Nonenum_beams: int | None = Nonepenalty_alpha: float | None = Nonetemperature: float | None = Nonetop_k: int | None = Nonetop_p: float | None = Nonetypical_p: float | None = Noneuse_cache: bool | None = None )

文本生成过程的参数化

类 huggingface_hub.TextToAudioInput

< >

( inputs: strparameters: huggingface_hub.inference._generated.types.text_to_audio.TextToAudioParameters | None = None )

文本转音频推理的输入

类 huggingface_hub.TextToAudioOutput

< >

( audio: typing.Anysampling_rate: float )

文本转音频任务的推理输出

class huggingface_hub.TextToAudioParameters

< >

( generation_parameters: huggingface_hub.inference._generated.types.text_to_audio.TextToAudioGenerationParameters | None = None )

用于文本转音频的额外推理参数

text_to_image

类 huggingface_hub.TextToImageInput

< >

( inputs: strparameters: huggingface_hub.inference._generated.types.text_to_image.TextToImageParameters | None = None )

文本到图像推理的输入

类 huggingface_hub.TextToImageOutput

< >

( image: typing.Any )

文本到图像任务的推理输出

class huggingface_hub.TextToImageParameters

< >

( guidance_scale: float | None = Noneheight: int | None = Nonenegative_prompt: str | None = Nonenum_inference_steps: int | None = Nonescheduler: str | None = Noneseed: int | None = Nonewidth: int | None = None )

用于文本到图像的额外推理参数

text_to_speech

class huggingface_hub.TextToSpeechGenerationParameters

< >

( do_sample: bool | None = Noneearly_stopping: typing.Union[bool, ForwardRef('TextToSpeechEarlyStoppingEnum'), NoneType] = Noneepsilon_cutoff: float | None = Noneeta_cutoff: float | None = Nonemax_length: int | None = Nonemax_new_tokens: int | None = Nonemin_length: int | None = Nonemin_new_tokens: int | None = Nonenum_beam_groups: int | None = Nonenum_beams: int | None = Nonepenalty_alpha: float | None = Nonetemperature: float | None = Nonetop_k: int | None = Nonetop_p: float | None = Nonetypical_p: float | None = Noneuse_cache: bool | None = None )

文本生成过程的参数化

类 huggingface_hub.TextToSpeechInput

< >

( inputs: strparameters: huggingface_hub.inference._generated.types.text_to_speech.TextToSpeechParameters | None = None )

文本转语音推理的输入

类 huggingface_hub.TextToSpeechOutput

< >

( audio: typing.Anysampling_rate: float | None = None )

文本转语音任务的推理输出

class huggingface_hub.TextToSpeechParameters

< >

( generation_parameters: huggingface_hub.inference._generated.types.text_to_speech.TextToSpeechGenerationParameters | None = None )

用于文本转语音的额外推理参数

text_to_video

类 huggingface_hub.TextToVideoInput

< >

( inputs: strparameters: huggingface_hub.inference._generated.types.text_to_video.TextToVideoParameters | None = None )

文本生成视频推理的输入

类 huggingface_hub.TextToVideoOutput

< >

( video: typing.Any )

文本生成视频任务的推理输出

class huggingface_hub.TextToVideoParameters

< >

( guidance_scale: float | None = Nonenegative_prompt: list[str] | None = Nonenum_frames: float | None = Nonenum_inference_steps: int | None = Noneseed: int | None = None )

用于文本生成视频的额外推理参数

token_classification

类 huggingface_hub.TokenClassificationInput

< >

( inputs: strparameters: huggingface_hub.inference._generated.types.token_classification.TokenClassificationParameters | None = None )

用于 Token Classification 推理的输入

类 huggingface_hub.TokenClassificationOutputElement

< >

( end: intscore: floatstart: intword: strentity: str | None = Noneentity_group: str | None = None )

Token Classification 任务的推理输出

class huggingface_hub.TokenClassificationParameters

< >

( aggregation_strategy: typing.Optional[ForwardRef('TokenClassificationAggregationStrategy')] = Noneignore_labels: list[str] | None = Nonestride: int | None = None )

用于 Token Classification 的额外推理参数

translation

类 huggingface_hub.TranslationInput

< >

( inputs: strparameters: huggingface_hub.inference._generated.types.translation.TranslationParameters | None = None )

翻译推理的输入

类 huggingface_hub.TranslationOutput

< >

( translation_text: str )

翻译任务的推理输出

类 huggingface_hub.TranslationParameters

< >

( clean_up_tokenization_spaces: bool | None = Nonegenerate_parameters: dict[str, typing.Any] | None = Nonesrc_lang: str | None = Nonetgt_lang: str | None = Nonetruncation: typing.Optional[ForwardRef('TranslationTruncationStrategy')] = None )

用于翻译的额外推理参数

video_classification

类 huggingface_hub.VideoClassificationInput

< >

( inputs: typing.Anyparameters: huggingface_hub.inference._generated.types.video_classification.VideoClassificationParameters | None = None )

视频分类推理的输入

类 huggingface_hub.VideoClassificationOutputElement

< >

( label: strscore: float )

视频分类任务的推理输出

类 huggingface_hub.VideoClassificationParameters

< >

( frame_sampling_rate: int | None = Nonefunction_to_apply: typing.Optional[ForwardRef('VideoClassificationOutputTransform')] = Nonenum_frames: int | None = Nonetop_k: int | None = None )

用于视频分类的额外推理参数

visual_question_answering

类 huggingface_hub.VisualQuestionAnsweringInput

< >

( inputs: VisualQuestionAnsweringInputDataparameters: huggingface_hub.inference._generated.types.visual_question_answering.VisualQuestionAnsweringParameters | None = None )

视觉问答推理的输入

类 huggingface_hub.VisualQuestionAnsweringInputData

< >

( image: typing.Anyquestion: str )

一个（图像，问题）对用于回答

类 huggingface_hub.VisualQuestionAnsweringOutputElement

< >

( score: floatanswer: str | None = None )

视觉问答任务的推理输出

类 huggingface_hub.VisualQuestionAnsweringParameters

< >

( top_k: int | None = None )

用于视觉问答的额外推理参数

zero_shot_classification

类 huggingface_hub.ZeroShotClassificationInput

< >

( inputs: strparameters: ZeroShotClassificationParameters )

Zero Shot Classification 推理的输入

类 huggingface_hub.ZeroShotClassificationOutputElement

< >

( label: strscore: float )

Zero Shot Classification 任务的推理输出

类 huggingface_hub.ZeroShotClassificationParameters

< >

( candidate_labels: listhypothesis_template: str | None = Nonemulti_label: bool | None = None )

Zero Shot Classification 的额外推理参数

zero_shot_image_classification

类 huggingface_hub.ZeroShotImageClassificationInput

< >

( inputs: strparameters: ZeroShotImageClassificationParameters )

Zero Shot Image Classification 推理的输入

类 huggingface_hub.ZeroShotImageClassificationOutputElement

< >

( label: strscore: float )

Zero Shot Image Classification 任务的推理输出

类 huggingface_hub.ZeroShotImageClassificationParameters

< >

( candidate_labels: listhypothesis_template: str | None = None )

Zero Shot Image Classification 的额外推理参数

zero_shot_object_detection

类 huggingface_hub.ZeroShotObjectDetectionBoundingBox

< >

( xmax: intxmin: intymax: intymin: int )

预测的边界框。坐标相对于输入图像的左上角。

类 huggingface_hub.ZeroShotObjectDetectionInput

< >

( inputs: strparameters: ZeroShotObjectDetectionParameters )

Zero Shot Object Detection 推理的输入

类 huggingface_hub.ZeroShotObjectDetectionOutputElement

< >

( box: ZeroShotObjectDetectionBoundingBoxlabel: strscore: float )

Zero Shot Object Detection 任务的推理输出

类 huggingface_hub.ZeroShotObjectDetectionParameters

< >

( candidate_labels: list )

Zero Shot Object Detection 的额外推理参数

Update on GitHub

←Mixins 与序列化方法 推理客户端→