All - Activity - prasanth55555 - Rasa Community Forum

Large Memory & Time Utilization for Loading a Model

Hi Guys, Hope you all are doing well, I have some queries. I am working on building a model which needs the NLU alone, the records which I use in the lookup table for slot identification has about 25-30 Lakh data in it. I also have a Custom component.

My Pipelines are:

language: "en"
pipeline:
  - name: "WhitespaceTokenizer"
  - name: "RegexFeaturizer"
  - name: "deepPavlov.DeepPavlov"
  - name: "CRFEntityExtractor"

@capgos17 lets say for example I have two intents namely:

intent: listVideoIntent

intent: listAudioIntent

I have these kinds of similar intents. I can't have these intents in the same intent.

Hi guys, I have a doubt about my training data. I am to create a NLU with multiple intents and I have a large amount of data for training so I prefer this config.

language: "en"
pipeline:
  - name: "WhitespaceTokenizer"
  - name: "RegexFeaturizer"
  - name: "deepPavlov.DeepPavlov"
  - name: "CRFEntityExtractor"

@stephens thanks for your reply, I am using the below config file.

language: "en"
pipeline:
  - name: "WhitespaceTokenizer"
  - name: "RegexFeaturizer"
  - name: "deepPavlov.DeepPavlov"
  - name: "CRFEntityExtractor"
  - features: [["low", "title", "upper"], ["bias", "low", "prefix5", "prefix2", "suffix5", "suffix3"]]

Question

{
  "text": "search for the book alice by james"
}

Response:

{
  "intent": {
    "name": "SearchIntent",
    "confidence": 0.9857004881
  },
  "entities": [
    {
      "start": 29,
      "end": 34,
      "value": "james",
      "entity": "Person",
      "confidence": 0.9153108498,
      "extractor": "CRFEntityExtractor"
    }
  ],
  "intent_ranking": [
    ...
  ]
}

to search for a book:

@fuih the issue is that I need to capture both entities: book name and author name dynamically. The regex needs to be improved.

Final Note

Is there any way to recognize dynamic values that are being replaced by synonyms? My NLU correctly recognizes the synonym value but I need to log the original value.