im a freshman on rasa (chatbot). i am using rasa Version 2.2 with botfront for admin panel. my project was deployed and publish in development area (restricted user on my work only) in 3 last days.
my question :
- when i train rasa in botfront apps and see the log on server, he will train based on session user. is it like that ? here my log on server
2023-02-08 06:23:36 DEBUG rasa_addons.core.channels.webchat - User BMkLfyfz-39wX1NPAAit connected to socketIO endpoint.
2023-02-08 06:23:42 DEBUG rasa.utils.tensorflow.models - Finished building tensorflow prediction graph.
2023-02-08 06:23:42 DEBUG rasa.core.nlg.generator - Instantiated NLG to 'GraphQLNaturalLanguageGenerator'.
2023-02-08 06:23:42 DEBUG rasa.server -Successfully loaded model 'model-5pcXs4w7mJuwRH5C7.tar.gz'.
Epochs: 18%|█▊ | 9/50 [04:05<17:36, 25.78s/it, t_loss=1.653, loss=0.690, acc=0.998]2023-02-08 06:24:00 DEBUG rasa_addons.core.channels.webchat - User tRpkKF7BlMZm8PNPAAip disconnected from socketIO endpoint.
2023-02-08 06:24:03 DEBUG rasa_addons.core.channels.webchat - User l4aoVZpt-xzfb6pvAAiv connected to socketIO endpoint.
2023-02-08 06:24:04 DEBUG rasa_addons.core.channels.graphql - fetching endpoints and credentials at http://botfront-app:3000/graphql
2023-02-08 06:24:04 DEBUG rasa_addons.core.channels.webchat - User l4aoVZpt-xzfb6pvAAiv connected to socketIO endpoint.
Epochs: 22%|██▏ | 11/50 [04:36<13:15, 20.39s/it, t_loss=1.371, loss=0.427, acc=0.998]2023-02-08 06:24:37 DEBUG rasa_addons.core.channels.webchat - User f2Jmi43Qv656OJTAAAix connected to socketIO endpoint.
2023-02-08 06:24:37 DEBUG rasa_addons.core.channels.graphql - fetching endpoints and credentials at http://botfront-app:3000/graphql
Epochs: 24%|██▍ | 12/50 [04:50<11:49, 18.67s/it, t_loss=1.275, loss=0.343, acc=0.999]2023-02-08 06:24:37 DEBUG rasa_addons.core.channels.webchat - User f2Jmi43Qv656OJTAAAix connected to socketIO endpoint.
Epochs: 30%|███ | 15/50 [05:34<09:10, 15.73s/it, t_loss=1.094, loss=0.194, acc=0.999]2023-02-08 06:25:23 DEBUG rasa_addons.core.channels.webchat - User l4aoVZpt-xzfb6pvAAiv disconnected from socketIO endpoint.
and many more epoch process on different user session. on the log, he's already done train and load the model.
- if yes, how can i train this core without based on session server ?
my expect, train in the same time when the endpoint rasa is published and accessing from user.