-
Notifications
You must be signed in to change notification settings - Fork 2.5k
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
* Create FAQ.md * Update FAQ.md * Update FAQ.md * Update FAQ.md * Update FAQ.md * Update FAQ.md * Update FAQ.md
- Loading branch information
Showing
1 changed file
with
60 additions
and
0 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,60 @@ | ||
# Frequently Asked Questions | ||
|
||
### What are the primary scenarios supported by Orchestrator? | ||
Primary scenario currently supported is routing/dispatching (intent only). We are working on supporting entities extraction for future releases. | ||
|
||
|
||
### Can Orchestrator be used from Composer? | ||
Yes. Follow documentation [here][13] to use Orchestrator as recognizer in Composer. | ||
|
||
### Can you use both Orchestrator and LUIS in the same bot? | ||
Yes, definitely. You could use Orchestrator (OrchestratorRecognizer) to route incoming utterances to the right child applications (i.e., LUISRecognizer, QnAMakerRecognizer). | ||
|
||
|
||
### What are the guidelines for when Orchestrator correctly predicts incoming utterance for sub intent (app/skill) but the sub app fails to return the right intent? | ||
We recommend tuning the sub app models (i.e. LUIS/QnAMaker/Orchestrator) using the recommended best practices for the app. | ||
- To improve LUIS model, follow instructions [here][1]. | ||
- To improve QnAMaker kb, follow guidelines [here][2]. | ||
- To improve Orchestrator, follow guidelines [here][3]. | ||
|
||
|
||
### Which languages will be supported in Orchestrator multilingual models? | ||
Orchestrator multilingual models currently supports 8 languages (EN, ES, DE, FR, IT, JA, PT, and ZH). See [Orchestrator base models][14] for more information on specific models. We are constantly looking to expand the number of supported languages and if a language is not yet supported, Azure translation could be used to translate incoming utterance to English first before passing it to Orchestrator. | ||
|
||
### Does Orchestrator replace LUIS Dispatch? | ||
LUIS based [Dispatch][5] is on the path to be deprecated, although the exact timeline is not yet finalized. Now is a good time to start evaluating Orchestrator as an alternative to Dispatch. | ||
|
||
|
||
### Does Orchestrator support entity extraction? | ||
Not at the moment, but we are working on it for future releases. | ||
|
||
|
||
### What is the guidelines in choosing Orchestrator vs LUIS? | ||
We recommend evaluating each solution against your data and see which performs better. Both LUIS based [Dispatch CLI][5] and [BF Orchestrator CLI][6] come with a command to evaluate training and test data. | ||
|
||
### Could you combine Orchestrator to predict intent and LUIS for entity? | ||
Unfortunately not. However, Orchestrator Recognizer could take external entity recognizers. Look [here][7] for supported entity recognizers. | ||
|
||
### How do you migrate from Dispatch solution to Orchestrator solution? | ||
[BF Orchestrator CLI][6] comes with commands similar to [Dispatch CLI][5] commands. bf orchestrator:create command also takes existing .dispatch file generated from Dispatch CLI for easy migration. See [Dispatch Migration][12] documentation for more information. | ||
|
||
### References | ||
* [Orchestrator][8] | ||
* [Composer][9] | ||
* [LUIS][10] | ||
* [QnAMaker][11] | ||
|
||
[1]:https://docs.microsoft.com/en-us/azure/cognitive-services/luis/luis-how-to-use-dashboard | ||
[2]:https://docs.microsoft.com/en-us/azure/cognitive-services/qnamaker/how-to/test-knowledge-base?tabs=v1 | ||
[3]:https://github.com/microsoft/BotBuilder-Samples/blob/main/experimental/orchestrator/CLI/ModelTuning/README.md | ||
[4]:https://github.com/pytorch/fairseq/tree/master/examples/xlmr#introduction | ||
[5]:https://github.com/microsoft/botbuilder-tools/tree/master/packages/Dispatch | ||
[6]:https://github.com/microsoft/botframework-cli/tree/main/packages/Orchestrator | ||
[7]:https://github.com/microsoft/botbuilder-dotnet/tree/main/libraries/Microsoft.Bot.Builder.Dialogs.Adaptive/Recognizers/EntityRecognizers | ||
[8]:https://aka.ms/bf-orchestrator | ||
[9]:https://github.com/microsoft/BotFramework-Composer | ||
[10]:https://docs.microsoft.com/en-us/azure/cognitive-services/luis/what-is-luis | ||
[11]:https://docs.microsoft.com/en-us/azure/cognitive-services/qnamaker/overview/overview | ||
[12]:https://github.com/microsoft/botframework-sdk/blob/tiens-patch-1/Orchestrator/docs/DispatchMigrationExample.md | ||
[13]:https://github.com/microsoft/BotFramework-Composer/blob/main/docs/preview%20features/orchestrator.md | ||
[14]:https://github.com/microsoft/botframework-sdk/blob/main/Orchestrator/docs/NLRModels.md |