I’m unsuccesfully trying to use the Two-Stage Fallback policy.
As you can see in this example:
Instead of asking the user of reformulating the message, it just asks again if the intent was correct and then it never triggers the default fallback.
My config file is the following:
language: en pipeline: supervised_embeddings policies: - name: TwoStageFallbackPolicy nlu_threshold: 0.3 ambiguity_threshold: 0.1 core_threshold: 0.3 fallback_core_action_name: "action_default_fallback" fallback_nlu_action_name: "action_default_fallback" deny_suggestion_intent_name: "out_of_scope" - name: MemoizationPolicy - name: MappingPolicy - name: EmbeddingPolicy max_history: 10 batch_strategy: balanced epochs: 5 random_seed: 1234 evaluate_on_num_examples: 0
Thank you! Tiziano