LUIS 模型和密钥的限制Limits for your LUIS model and keys

LUIS 有几个限制区。LUIS has several limit areas. 第一个是模型限制,它可控制 LUIS 中的意向、实体和功能。The first is the model limit, which controls intents, entities, and features in LUIS. 第二个是基于密钥类型的配额限制The second area is quota limits based on key type. 限制的第三个区域是用于控制 LUIS 网站的键盘组合A third area of limits is the keyboard combination for controlling the LUIS website. 第四个是 LUIS 创作网站和 LUIS 终结点 API 之间的区域映射A fourth area is the region mapping between the LUIS authoring website and the LUIS endpoint APIs.

模型限制Model limits

如果应用超出 LUIS 模型限制,请考虑使用 LUIS 调度应用或使用 LUIS 容器If your app exceeds the LUIS model limits, consider using a LUIS dispatch app or using a LUIS container.

区域Area 限制Limit
应用名称App name *默认最大字符长度*Default character max
应用程序Applications 每个 Azure 创作资源 500 个应用程序500 applications per Azure authoring resource
批处理测试Batch testing 10 个数据集,每个数据集 1000 条话语10 datasets, 1000 utterances per dataset
显式列表Explicit list 每个应用程序 50 个实体50 per application
外部实体External entities 无限制no limits
意向Intents 每个应用程序 500 个意向:499 个自定义意向,以及必需的“无”意向。500 per application: 499 custom intents, and the required None intent.
基于调度的应用程序具有相应的 500 个调度源。Dispatch-based application has corresponding 500 dispatch sources.
列表实体List entities 父级:50 个,子级:20,000 项。Parent: 50, child: 20,000 items. 规范名称是 *默认最大字符长度。同义词值没有长度限制。Canonical name is *default character max. Synonym values have no length restriction.
机器学习实体 + 角色machine-learning entities + roles:
复合、composite,
简单、simple,
实体角色entity role
限制为 100 个父实体或 330 个实体,以用户先达到的限制为准。A limit of either 100 parent entities or 330 entities, whichever limit the user hits first. 就此限制来说,一个角色计为一个实体。A role counts as an entity for the purpose of this limit. 例如,一个复合实体包含一个简单实体,并且有 2 个角色:1 个复合实体 + 1 个简单实体 + 2 个角色 = 4 个实体(最多 330 个实体)。An example is a composite with a simple entity, which has 2 roles is: 1 composite + 1 simple + 2 roles = 4 of the 330 entities.
子实体最多可以嵌套 5 级。Subentities can be nested up to 5 levels.
作为特征的模型Model as a feature 可用作特定模型的特征的模型的最大数量为 10 个模型。Maximum number of models that can be used as a feature to a specific model to be 10 models. 可用作特定模型的特征的短语列表的最大数量为 10 个短语列表。The maximum number of phrase lists used as a feature for a specific model to be 10 phrase lists.
预览 - 动态列表实体Preview - Dynamic list entities 2 个列表,其中的每个查询预测终结点请求约 1k 个实体2 lists of ~1k per query prediction endpoint request
模式Patterns 每个应用程序 500 个模式。500 patterns per application.
模式的最大长度为 400 个字符。Maximum length of pattern is 400 characters.
每个模式 3 个 Pattern.any 实体3 Pattern.any entities per pattern
模式中最多 2 个嵌套可选文本Maximum of 2 nested optional texts in pattern
Pattern.anyPattern.any 每个应用程序 100 个,每个模式 3 个 pattern.any 实体100 per application, 3 pattern.any entities per pattern
短语列表Phrase list 500 个短语列表。500 phrase lists. 10 个全局短语列表(由于存在作为特征的模型的限制)。10 global phrase lists due to the model as a feature limit. 不可互换的短语列表最多有 5,000 个短语。Non-interchangeable phrase list has max of 5,000 phrases. 可互换的短语列表最多有 50,000 个短语。Interchangeable phrase list has max of 50,000 phrases. 每个应用程序的最大短语总数为 500,000 个短语。Maximum number of total phrases per application of 500,000 phrases.
正则表达式实体Regular expression entities 20 个实体20 entities
每个正则表达式实体模式500 character max. 最多 500 个字符per regular expression entity pattern
角色Roles 每个应用程序 300 个角色。300 roles per application. 每个实体 10 个角色10 roles per entity
话语Utterance 500 个字符500 characters

如果文本长度超过此字符限制,则需要在输入 LUIS 之前对话语进行分段,并且每个分段都会收到单独的意向响应。If you have text longer than this character limit, you need to segment the utterance prior to input to LUIS and you will receive individual intent responses per segment. 你可以处理一些明显的中断,比如标点符号和言语中的长停顿。There are obvious breaks you can work with, such as punctuation marks and long pauses in speech.
言语示例Utterance examples 每个应用程序 15,000 条 - 对每个意向的话语数量没有限制15,000 per application - there is no limit on the number of utterances per intent

如果需要使用更多示例来定型应用程序,请使用 dispatch 模型方法。If you need to train the application with more examples, use a dispatch model approach. 可以使用一个或多个意向定型单个 LUIS 应用(称为父调度应用的子应用),然后定型调度应用,该应用从每个 LUIS 子应用的言语中采样,以将预测请求定向到正确的子应用。You train individual LUIS apps (known as child apps to the parent dispatch app) with one or more intents and then train a dispatch app that samples from each child LUIS app's utterances to direct the prediction request to the correct child app.
版本Versions 每个应用程序 100 个版本100 versions per application
版本名称Version name 128 个字符128 characters

*默认最大字符长度为 50 个字符。*Default character max is 50 characters.

名称唯一性Name uniqueness

与相同级别的其他对象进行比较时,对象名称必须独一无二。Object names must be unique when compared to other objects of the same level.

对象Objects 限制Restrictions
意向、实体Intent, entity 所有意向和实体名称在一个应用版本中必须独一无二。All intent and entity names must be unique in a version of an app.
ML 实体组件ML entity components 对于同一级别的组件,所有机器学习实体组件(子实体)都必须在该实体内独一无二。All machine-learning entity components (child entities) must be unique, within that entity for components at the same level.
功能Features 所有已命名特征(例如短语列表)在一个应用版本中必须独一无二。All named features, such as phrase lists, must be unique within a version of an app.
实体角色Entity roles 当实体或实体组件上的所有角色处于同一实体级别(父级、子级、孙级等)时,它们必须独一无二。All roles on an entity or entity component must be unique when they are at the same entity level (parent, child, grandchild, etc.).

对象命名Object naming

请勿在下列名称中使用下列字符。Do not use the following characters in the following names.

ObjectObject 排除字符Exclude characters
意向、实体和角色名称Intent, entity, and role names :
$
&
版本名称Version name \
/
:
?
&
=
*
+
(
)
%
@
$
~
!
#

资源使用情况和限制Resource usage and limits

语言理解有单独的资源,一种用于创作,一种用于查询预测终结点。Language Understand has separate resources, one type for authoring, and one type for querying the prediction endpoint. 若要详细了解密钥类型之间的差异,请参阅 LUIS 中的创作密钥和查询预测终结点密钥To learn more about the differences between key types, see Authoring and query prediction endpoint keys in LUIS.

创作资源限制Authoring resource limits

筛选 Azure 门户中的资源时,请使用 kind LUIS.AuthoringUse the kind, LUIS.Authoring, when filtering resources in the Azure portal. LUIS 的限制是每个 Azure 创作资源 500 个应用程序。LUIS limits 500 applications per Azure authoring resource.

创作资源Authoring resource 创作 TPSAuthoring TPS
初学者Starter 1 百万/月,5/秒1 million/month, 5/second
F0 - 免费层F0 - Free tier 1 百万/月,5/秒1 million/month, 5/second
  • TPS = 每秒事务数TPS = Transactions per second

详细了解定价。Learn more about pricing.

查询预测资源限制Query prediction resource limits

筛选 Azure 门户中的资源时,请使用 kind LUIS。在运行时上使用的 LUIS 查询预测终结点资源仅对终结点查询有效。Use the kind, LUIS, when filtering resources in the Azure portal.The LUIS query prediction endpoint resource, used on the runtime, is only valid for endpoint queries.

查询预测资源Query Prediction resource 查询 TPSQuery TPS
F0 - 免费层F0 - Free tier 1 万/月,5/秒10 thousand/month, 5/second
S0 - 标准层S0 - Standard tier 50/秒50/second

情绪分析Sentiment analysis

提供了情绪分析集成(其中提供情绪信息),无需其他 Azure 资源。Sentiment analysis integration, which provides sentiment information, is provided without requiring another Azure resource.

语音集成Speech integration

语音集成每单位成本提供 1 千个终结点请求。Speech integration provides 1 thousand endpoint requests per unit cost.

详细了解定价。Learn more about pricing.

键盘控件Keyboard controls

键盘输入Keyboard input 说明Description
Control+EControl+E 切换话语列表中的令牌和实体switches between tokens and entities on utterances list

网站登录时间段Website sign-in time period

登录访问权限为 60 分钟。Your sign-in access is for 60 minutes. 在此时间段后,会出现此错误。After this time period, you will get this error. 需重新登录。You need to sign in again.