Calling POST /train with a customized pipeline

Is there a format to provide a custom pipeline to the /train endpoint?

I’ve tried several variations including both yml and json formatted and I can’t seem to get a successful response when providing a custom pipeline.

The endpoint works fine when I use the preconfigured templates, but unfortunately I can’t just use the spacy_sklearn template. my project has been using a customized pipeline since early versions of RASA and I need to keep those components consistent to preserve expected behavior from our bot.

Yep, just like this :

["name": "nlp_spacy"}, { "name": "tokenizer_whitespace"}, { "name": "ner_crf"}, { "name": "ner_synonyms"}, { "name": "intent_featurizer_count_vectors"}, { "name": "intent_classifier_tensorflow_embedding"}]