Segmentation fault rasa train on enabling Intentless Policy - Rasa CALM - Rasa Community Forum

Segmentation fault rasa train on enabling Intentless Policy

post by geeta.m.desai on Feb 27, 2024

When I turn on IntentlessPolicy as below

• name: rasa_plus.ml.IntentlessPolicy nlu_abstention_threshold: 0.2

I get

2024-02-27 18:58:29 INFO rasa.engine.training.hooks - Starting to train component ‘IntentlessPolicy’.  
2024-02-27 18:58:31 INFO faiss.loader - Loading faiss with AVX2 support.  
2024-02-27 18:58:31 INFO faiss.loader - Successfully loaded faiss with AVX2 support.  
zsh: segmentation fault rasa train

Kindly please help me to debug and solve this issue
Thanks,
Geeta

post by Arjaan on Feb 27, 2024

Hi @geeta.m.desai,

Can you provide more detail that would allow us to reproduce your issue?
Perhaps you have your model in a Github repo so we can have a look?

post by geeta.m.desai on Feb 27, 2024

Hi Arjaan,

Thank you for responding back. Please find my project here https://github.com/geeta-m-desai/rasa-calm/blob/main/config.yml

Earlier I had a lot of code, I removed everything and simplified to see if it fixes the problem but it did not help

I have tried creating a new environment and other such techniques but did not help
Thanks,

post by jtrasa on Mar 1, 2024

Hi Geeta, I cloned your repo and was able to train successfully with just rasa train. This is with python 3.9.16 and rasa-plus 3.7.8. What versions are you using?

post by geeta.m.desai on Mar 1, 2024

Links:

post by jtrasa on Mar 2, 2024

Could you paste the output of the command rasa --version? This is a good start to debugging.

post by geeta.m.desai on Mar 2, 2024

(venv) geetadesai@MacBook-Pro rasa-calm % rasa --version
Rasa Version : 3.7.7
Minimum Compatible Version: 3.7.0
Rasa SDK Version : 3.7.1
Python Version : 3.10.10
Operating System : macOS-14.3.1-x86_64-i386-64bit
Python Path : /Users/geetadesai/rasa-calm/venv/bin/python3

post by jtrasa on Mar 4, 2024

Here is my rasa --version:

geeta-rasa-calm % rasa --version
Rasa Version      :         3.7.8
Minimum Compatible Version: 3.7.0
Rasa SDK Version  :         3.7.1
Python Version    :         3.10.13
Operating System  :         macOS-14.3.1-x86_64-i386-64bit
Python Path       :         /Users/john/src/github.com/trammell/bots/geeta-rasa-calm/.direnv/python-3.10/bin/python
Rasa Plus Version  :         3.7.8

post by geeta.m.desai on Mar 5, 2024

But I am using CALM and I want to route all the unidentified queries to this new CustomKnowledgeAnswerCommand

Is it possible to specify a specific action as a fallback when using CALM ?

I had tried below in flows.yml

fallback:
  description: fallback action
  name: fallback
  steps:
      - action: action_custom_knowledge_answer

post by geeta.m.desai on Mar 9, 2024

Thank you. I have moved past these issues for building my voice bot using Rasa-CALM.

I have functionality for reminders. I have configured as per documentation. But I see warning as (venv) geetadesai@MacBook-Pro rasa-calm % rasa data validate

2024-03-10 04:02:06 INFO rasa.validator - Validating intents…
/User/geetadesai/rasa-calm/venv/lib/python3.10/site-packages/rasa/shared/utils/io.py:99: UserWarning: The intent ‘EXTERNAL_reminder_callback’ is not used in any story, rule, or flow.

post by geeta.m.desai on Mar 10, 2024

I’m building a multilingual voice bot with CALM using a single model and a language translation service, and it’s slow. Is a multi-model architecture generally better for performance in this scenario or are there any alternatives?
Here is a brief detail of my flow:

1. STT module processes the speech.
2. Take the output of the STT module (text) and send it to the REST channel of your Rasa server.
3. Use Translation service to translate to user language.
4. Take the REST response (text) and send it to the TTS module.

How can I reduce lag in my Rasa API responses? Would a multi model be a suitable solution for overall speed improvement?

post by jtrasa on Mar 15, 2024

I really like to share all problems and especially in the entire new journey that I am on, it is valuable to share. But this time it was an issue with GCP firewall hence I considered it out of scope here. But will consider sharing. Thanks, Geeta

post by geeta.m.desai on Mar 18, 2024

I’m building a multilingual voice bot with CALM using a single model and a language translation service, and it’s slow. Is a multi-model architecture generally better for performance in this scenario or are there any alternatives? Here is brief detail of my flow:

And time taken for each is as below:

• Conversion took -1(STT) -->: 3.45 seconds
• Conversion took -2(Rasa Response) -->: 24.54 seconds
  response_content → Warm greetings from MitramCares! I’m your friendly helper. How can I make things easier for you today?
• Conversion took -3 (Translation)–>: 0.76 seconds
• Conversion took -4(TTS) -->: 0.99 seconds

How can I reduce lag in my Rasa API responses? Would a multi model be a suitable solution for overall speed improvement?

post by jtrasa on Mar 19, 2024

Rasa Pro 3.7 still supports NLU functionality, so the examples in the documentation are a good starting point.

post by geeta.m.desai on Mar 20, 2024

Thank you, will continue with NLU style for reminders.