Number selection does not seem to work

I have the following NLU

version: "2.0"
nlu:
- intent: UserSelectedNumber
  examples: |
    - [one]{"entity": "number"}
    - I would say [two]{"entity": "number"}
    - I think [three]{"entity": "number"}
    - I think it's [three]{"entity": "number"}
    - I reckon [ten]{"entity": "number"}
    - I reckon it's a [ten]{"entity": "number"}
    - definitely [six]{"entity": "number"}

- lookup: number
  examples: |
    - one
    - two
    - three
    - four
    - five
    - six
    - seven
    - eight
    - nine
    - ten

However classification seems to fail (either it doesn’t identify entity or misses it all together example: “five”

{
  "text": "five",
  "intent": {
    "id": -6058370058712717852,
    "name": "UserSelectedNumber",
    "confidence": 0.7820941209793091
  },
  "entities": [],
  "intent_ranking": [
    {
      "id": -6058370058712717852,
      "name": "UserSelectedNumber",
      "confidence": 0.7820941209793091
    },
    {
      "id": -7727331658891479453,
      "name": "UserRespondedWithClarification",
      "confidence": 0.0612444132566452
    },
    {
      "id": -2418204329415073218,
      "name": "greet",
      "confidence": 0.05769157037138939
    },
    {
      "id": -9174561335986908344,
      "name": "UserRespondedWithPositiveInterruption",
      "confidence": 0.020679784938693047
    },
    {
      "id": 7498480174661378028,
      "name": "deny",
      "confidence": 0.018534181639552116
    },
    {
      "id": 8539080034510111850,
      "name": "UserRespondedWithMumble",
      "confidence": 0.01603601686656475
    },
    {
      "id": 8883814552498850386,
      "name": "affirm",
      "confidence": 0.014628269709646702
    },
    {
      "id": 7961907735163220506,
      "name": "UserRespondedWithRepeatQuestion",
      "confidence": 0.010131706483662128
    },
    {
      "id": 4592032478428692444,
      "name": "UserRespondedWithChitChat",
      "confidence": 0.009761279448866844
    },
    {
      "id": 971804410359964896,
      "name": "goodbye",
      "confidence": 0.009198571555316448
    }
  ]
}

but “Two” works

two
{
  "text": "two",
  "intent": {
    "id": -6058370058712717852,
    "name": "UserSelectedNumber",
    "confidence": 0.9632925391197205
  },
  "entities": [
    {
      "entity": "number",
      "start": 0,
      "end": 3,
      "confidence_entity": 0.9323317408561707,
      "value": "two",
      "extractor": "DIETClassifier"
    }
  ],
  "intent_ranking": [
    {
      "id": -6058370058712717852,
      "name": "UserSelectedNumber",
      "confidence": 0.9632925391197205
    },
    {
      "id": -7727331658891479453,
      "name": "UserRespondedWithClarification",
      "confidence": 0.014467975124716759
    },
    {
      "id": 4592032478428692444,
      "name": "UserRespondedWithChitChat",
      "confidence": 0.004221928305923939
    },
    {
      "id": -9174561335986908344,
      "name": "UserRespondedWithPositiveInterruption",
      "confidence": 0.00361985107883811
    },
    {
      "id": 7498480174661378028,
      "name": "deny",
      "confidence": 0.003487028181552887
    },
    {
      "id": -2418204329415073218,
      "name": "greet",
      "confidence": 0.0033619117457419634
    },
    {
      "id": 8539080034510111850,
      "name": "UserRespondedWithMumble",
      "confidence": 0.0024840321857482195
    },
    {
      "id": 971804410359964896,
      "name": "goodbye",
      "confidence": 0.0020540303084999323
    },
    {
      "id": -7142475470372166967,
      "name": "UserRespondedWithNegativeInterruption",
      "confidence": 0.0015963840996846557
    },
    {
      "id": 7961907735163220506,
      "name": "UserRespondedWithRepeatQuestion",
      "confidence": 0.0014144022716209292
    }
  ]
}

here is my config

# Configuration for Rasa NLU.
# https://rasa.com/docs/rasa/nlu/components/
language: en
pipeline:
  - name: "WhitespaceTokenizer"
  - name: "RegexFeaturizer"
  - name: "LexicalSyntacticFeaturizer"
  - name: "CountVectorsFeaturizer"
  - name: "CountVectorsFeaturizer"
    analyzer: "word"
    min_ngram: 1
    max_ngram: 4
  - name: "DIETClassifier"
    epochs: 100
  - name: "EntitySynonymMapper"


# Configuration for Rasa Core.
# https://rasa.com/docs/rasa/core/policies/
policies:
  - name: RulePolicy

and domain

version: "2.0"

intents:
- UserRespondedWithMumble
- UserRespondedWithClarification
- UserRespondedWithChitChat
- UserRespondedWithHoldOffToCheck
- UserRespondedWithNegativeInterruption
- UserRespondedWithPositiveInterruption
- UserRespondedWithRepeatQuestion
- affirm
- deny
- greet
- goodbye
- UserSelectedNumber

entities:
- number

slots:
  number:
    type: unfeaturized

session_config:
  session_expiration_time: 0
  carry_over_slots_to_new_session: true

RASA version

Rasa Version     : 2.0.0rc2
Rasa SDK Version : 2.0.0rc1
Rasa X Version   : None
Python Version   : 3.7.6 (default, Jan  8 2020, 19:59:22) 
Operating System : Linux-4.19.104-microsoft-standard-x86_64-with-debian-bullseye-sid

Fyi, turns out I didn’t train the model long enough… cranking the number of epochs seems to have fixed it.

how much did you increase the epoch???

just picked 200 (original was 100)… My models are small so doubling the epochs doesn’t significantly increase training time.

thanks man