Hello,
I am having trouble training my NLU with language “de” model. I am getting error while training that “Model ‘de’ is not a linked spaCy model”. However spaCy info states “Models en_core_web_md, de_core_news_sm, de, en”. So, I am clueless what I am doing wrong here?
my config.yml:
language: de
pipeline:
- name: SpacyNLP
- name: SpacyTokenizer
- name: RegexFeaturizer
- name: CountVectorsFeaturizer
analyzer: char_wb
max_featurizes: 10000
max_ngram: 15
min_ngram: 2
- name: DucklingHTTPExtractor
dimensions:
- time
- duration
locale: de_DE
timezone: Europe/Berlin
url: http://duckling:8000
- name: DIETClassifier
epochs: 20
intent_tokenization_flag: true
intent_split_symbol: "+"
- name: EntitySynonymMapper
policies:
- name: TEDPolicy
batch_size: 50
epochs: 20
max_training_samples: 300
max_history: 5
- name: FallbackPolicy
fallback_action_name: action_default_fallback
nlu_threshold: 0.2
- name: MemoizationPolicy
max_history: 5
- name: FormPolicy
- name: MappingPolicy
Spacy info: ============================== Info about spaCy ==============================
spaCy version 2.2.3
Location /Users/Theresa/anaconda3/lib/python3.6/site-packages/spacy
Platform Darwin-19.3.0-x86_64-i386-64bit
Python version 3.6.2
Models en_core_web_md, de_core_news_sm, de, en
Error:
rasa.nlu.model.InvalidModelError: Model 'de' is not a linked spaCy model. Please download and/or link a spaCy model, e.g. by running:
python -m spacy download en_core_web_md
python -m spacy link en_core_web_md en
Training command:
rasa train --data data/carrots -c config.yml -d carrots/domain.yml --out models/carrots