From 3e5713a24fee3a4d54b5fe17e5dd97865db5ae21 Mon Sep 17 00:00:00 2001 From: Samarth Bhandari Date: Mon, 15 Oct 2018 11:18:50 -0700 Subject: [PATCH 1/2] Update Interactive learning documentation The documentation doesn't mention how to use a trained nlu model for intent classification with interactive learning. Although it is an intuitive command, people might not be able to figure it out so easily and just stay confused about why the intents from their model are not showing up. --- docs/interactive_learning.rst | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/docs/interactive_learning.rst b/docs/interactive_learning.rst index 5fb7b59b9a9..ced927dcfc3 100644 --- a/docs/interactive_learning.rst +++ b/docs/interactive_learning.rst @@ -29,7 +29,10 @@ Run the following to start interactive learning: python -m rasa_core.train \ --interactive -o models/dialogue \ -d domain.yml -s stories.md \ - --endpoints endpoints.yml + --endpoints endpoints.yml \ + --nlu models/current/nlu (OPTIONAL) + +To include an existing model to identify intents use --nlu models/current/nlu in the above command. Else interactive learning will use a default REGEX to intentify default intents from the user input text. The first command starts the action server (see :ref:`customactions`). From 243c70d5e5ac80a92602f3d4c53b135439199543 Mon Sep 17 00:00:00 2001 From: Samarth Bhandari Date: Tue, 16 Oct 2018 10:21:44 -0700 Subject: [PATCH 2/2] Update interactive_learning.rst Removed the flag. --- docs/interactive_learning.rst | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/docs/interactive_learning.rst b/docs/interactive_learning.rst index ced927dcfc3..bb47b98c5b3 100644 --- a/docs/interactive_learning.rst +++ b/docs/interactive_learning.rst @@ -29,8 +29,7 @@ Run the following to start interactive learning: python -m rasa_core.train \ --interactive -o models/dialogue \ -d domain.yml -s stories.md \ - --endpoints endpoints.yml \ - --nlu models/current/nlu (OPTIONAL) + --endpoints endpoints.yml To include an existing model to identify intents use --nlu models/current/nlu in the above command. Else interactive learning will use a default REGEX to intentify default intents from the user input text.