Train.py default: error: the following arguments are required: -c/--config

Getting Started with Rasa
1

I’m training my first rasa bot as per instructions in Quickstart. When I run the command: “python -m rasa_core.train -d domain.yml -s stories.md -o models/dialogue” I am getting the error “train.py default: error: the following arguments are required: -c/–config”

2

@Abir if you know something on this… :slight_smile:

3

you need to train using this command: python -m rasa_core.train -d domain.yml -s data/stories.md
-o models/dialogue -c default_config.yml

If you are using the training script, you must set the policies you would like the Core model to use in a YAML file i.e your config file.

For example:

policies:

  • name: “KerasPolicy” featurizer:
    • name: MaxHistoryTrackerFeaturizer max_history: 5 state_featurizer:
      • name: BinarySingleStateFeaturizer
  • name: “MemoizationPolicy” max_history: 5
  • name: “FallbackPolicy” nlu_threshold: 0.4 core_threshold: 0.3 fallback_action_name: “my_fallback_action”
  • name: “path.to.your.policy.class” arg1: “…”

for more details you can visit this link : Policies

4

@JiteshGaikwad has clarified it well , should solve your issue

5

@karandesaiii If you want to train the dialogues manually :

  1. Define train_dialogue

         def train_dialogue(domain_file=‘domain.yml’, 
                        model_path=’./models/dialogue’, 
                        training_data_file=’./data/nlu.md’):          

          agent = Agent(domain_file, policies=[MemoizationPolicy(), KerasPolicy()])
          data = agent.load_data(training_data_file)
          agent.train(data, batch_size=50, epochs=300, validation_split=0.2)
          agent.persist(model_path)
          return agent

  2. Use train_dialogue(). This will create the models/dialogue folder

  3. Option 1: run rasta server

    Option 2: Train NLU as in Quickstart and then run rasta server

6

Thanks guys, issue was resolved :slight_smile:

7

while training this command it gives the train.py:error

python -m rasa_nlu.train -c config.json

err1