Closed Primtek closed 6 years ago
@Primtek , I don't think there are restrictions on number of utterances and number of intent-checks. Are you facing any issues with your flow?
Yes, I am facing this problem.. Currently I use extension to extend the intent .. for example: Our utterance is in Bahasa Indonesia J
For example: [root@ai intents]# vi klaim.utterances.txt menyerempet keserempet diserempet ketabrak ditabrak nabrak ditubruk tubruk senggol disenggol
[root@ai intents]# vi klaim-ext001.utterances.txt menyerempet keserempet diserempet ketabrak ditabrak nabrak ditubruk tubruk senggol disenggol
I still don't get what's an issue.
In my case:
If I have klaim.utterance.txt as below:
[root@ai intents]# vi klaim.utterances.txt menyerempet keserempet diserempet ketabrak nabrak ditubruk tubruk senggol disenggol menyerempet keserempet diserempet ditabrak nabrak ditubruk tubruk senggol disenggol
and user types ‘ditabrak’ it would not be found .. it would jump to the ‘otherwise’ logic but if I separate it into 2 intents .. it would work.
I tried to adjust the “minimumConfidence” value .. but still no luck
[root@ai config]# vi nlu.json { "intentsDir": "./generated/intents", "entitiesDir": "./generated/entities", "provider": "native", "debugModeEnabled": true, "minimumConfidence": "0.3" }
I think this is not because of limits on utterances number, but because of issue in NLU-learning algorithm. @slvnperron , am I right?
Can you update your NLU module to the latest and see if that helps?
On Thu, May 31, 2018, 9:09 AM Primtek notifications@github.com wrote:
In my case:
If I have klaim.utterance.txt as below:
[root@ai intents]# vi klaim.utterances.txt
menyerempet
keserempet
diserempet
ketabrak
nabrak
ditubruk
tubruk
senggol
disenggol
menyerempet
keserempet
diserempet
ditabrak
nabrak
ditubruk
tubruk
senggol
disenggol
and user types ‘ditabrak’ it would not be found .. it would jump to the ‘otherwise’ logic
but if I separate it into 2 intents .. it would work.
I tried to adjust the “minimumConfidence” value .. but still no luck
[root@ai config]# vi nlu.json
{
"intentsDir": "./generated/intents",
"entitiesDir": "./generated/entities",
"provider": "native",
"debugModeEnabled": true,
"minimumConfidence": "0.3"
}
~
From: Alex Pletnov [mailto:notifications@github.com] Sent: 31 May 2018 19:57 To: botpress/botpress Cc: Primtek; Mention Subject: Re: [botpress/botpress] intent and utterance (#595)
I still don't get what's an issue.
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub < https://github.com/botpress/v12/issues/283> , or mute the thread < https://github.com/notifications/unsubscribe-auth/AlrhhZHlJvaVJ1EUMMnxcUnowajP5u5fks5t3-iAgaJpZM4UUugv> . < https://github.com/notifications/beacon/AlrhhV2vZPmTc0C7vrtbbDFDWw2a5Aefks5t3-iAgaJpZM4UUugv.gif>
Important Notice: The information contained in this e-mail and any attachment(s) is confidential and for the use of the named addressee(s) only. If you are not the intended recipient of this e-mail, do not duplicate, redistribute, or use it by any means. Please delete this e-mail and any attachment(s) and notify us immediately. Unauthorized use, reliance, disclosure or copying of the contents of this e-mail and any attachment(s), or any similar action, is strictly prohibited.
PT PRIMATAMA TEKNOLOGI SOLUSINDO reserves the right, to the extent permitted by applicable law, to retain, monitor and intercept e-mail messages both to and from its systems and please note that the integrity of e-mail across the internet cannot be guaranteed and PT PRIMATAMA TEKNOLOGI SOLUSINDO do not accept liability for loss or damage arising in any way from receipt, use or delay of this e-mail and any attachment(s) nor for improper or incomplete transmission of the information contained therein, by or to PT PRIMATAMA TEKNOLOGI SOLUSINDO
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/botpress/v12/issues/283, or mute the thread https://github.com/notifications/unsubscribe-auth/ABQStH0C61rWoLb4iJD9CXnn58X_455Yks5t3-uDgaJpZM4UUugv .
info: Loaded @botpress/nlu, version 10.11.2 it’s the latest right?
I created a simple project that only have 3 flows:
if user enter valid utterance then it will go to gotit otherwise it will go to unknown try to enter invalid utterance (ie. asda;sdk;la;lkasd;laksd;asd) then it also will go to gotit
I run bp 10.11.2
with this new fresh project, you also find the warning:
warn: DEPRECATION NOTICE â bp.umm is deprecated and will be removed in botpress@3.0
â Please see botpress/v12#282
and with this project, I also unable to get welcometext message for webchat.. i believe this project already exactly the same with the sample project on dev-bot I am stuck on this.. I did not find complete ref or doc for this.. you can find this project on https://github.com/Primtek/simple_but_fail
Ok, thanks for sharing sources. I'll have a look.
Thanks
Run it .. then you will find the problem of botpress/v12#282 --warn: DEPRECATION NOTICE â bp.umm is deprecated and will be removed in botpress@3.0
â Please see bp.renderers instead (#593)
And use webchat .. then you will find the problem of botpress/v12#284
that is a brand new fresh project, so I believe it exactly the same with dev-bot
thanks a lot in advance
Hi Good morning Alex, Do you have any update? ..
I updated to bp 10.13.4 but still facing this problem
Thanks
Hello @slvnperron ,
Do you have any update regarding to this issue? I have updated to 10.13.4 but the same problem still occurs.. I enter a text that does not exist in intent .. but it still accepted..
Again, it looks like there is a max. number of utterance in an intent because if I remove some utterance, it will work properly
Need your advice..
@Primtek , did you close the issue because it's no more the case?
I've checked some of the bots I was working on and some of the do have more 25 utterances and maybe more.
I don't think an issue is in number of utterances itself but in the machine-learning algorithm that fails to learn model based on some of the utterances you have if you put them into one group.
but what I don't understand is if I separate them into 2 intents then it works properly, any idea, what should I check? .. (native nlu)
My guess is that there are some utterances that "conflict" in AI-s logic when they are within same intent. But I'd rather wait for @slvnperron 's response.
Ok.. will wait for his response ..
Hello? Was the issue solved? I have same problem
@arbakag , please report your ploblem as a separate issue and don't forget to provide us with botpress
and @botpress/nlu
versions.
ok,thank you
Make sure these checkboxes are checked before raising an issue, thank you!
6.10
Please also fill in these fields:
what is the max of utterance per intents? .. and is there any max number of event.nlu,intent.is('myintent1') per flow? I am asking about this.. because I have a flow that control the intents, like below
Botpress version:
<<VERSION HERE>>
10,11.2
Channels: <
>
channel-web