Skip to content

Latest commit

 

History

History
58 lines (37 loc) · 3.58 KB

luis-how-to-train.md

File metadata and controls

58 lines (37 loc) · 3.58 KB
title titleSuffix description services author manager ms.service ms.component ms.topic ms.date ms.author
Train your LUIS version
Azure Cognitive Services
Training is the process of teaching your Language Understanding (LUIS) app version to improve its natural language understanding. Train your LUIS app after updates to the model such as adding, editing, labeling, or deleting entities, intents, or utterances.
cognitive-services
diberry
cgronlun
cognitive-services
language-understanding
article
10/18/2018
diberry

Train your LUIS app version

Training is the process of teaching your Language Understanding (LUIS) app to improve its natural language understanding. Train your LUIS app after updates to the model such as adding, editing, labeling, or deleting entities, intents, or utterances.

Training and testing an app is an iterative process. After you train your LUIS app, you test it with sample utterances to see if the intents and entities are recognized correctly. If they're not, make updates to the LUIS app, train, and test again.

Training is applied to the active version in the LUIS portal.

How to train interactively

To start the iterative process in the LUIS portal, you first need to train your LUIS app at least once. Make sure every intent has at least one utterance before training.

  1. Access your app by selecting its name on the My Apps page.

  2. In your app, select Train in the top panel.

  3. When training is complete, a green notification bar appears at the top of the browser.

Note

If you have one or more intents in your app that do not contain example utterances, you cannot train your app. Add utterances for all your intents. For more information, see Add example utterances.

Train with all data

Training uses a small percentage of negative sampling. If you want to use all data instead of the small negative sampling, use the Version settings API with the UseAllTrainingData set to true to turn off this feature.

Unnecessary training

You do not need to train after every single change. Training should be done after a group of changes are applied to the model, and the next step you want to do is to test or publish. If you do not need to test or publish, training isn't necessary.

Training with the REST APIs

Training in the LUIS portal is a single step of pressing the Train button. Training with the REST APIs is a two-step process. The first is to request training with HTTP POST. Then request the training status with HTTP Get.

In order to know when training is complete, you have to poll the status until all models are successfully trained.

Next steps