Just got a fresh install of Rasa and now when I run the shell, I get the following warning message after each input: c:\users\name\.conda\envs\rasa\lib\site-packages\rasa\nlu\extractors\crf_entity_extractor.py:590: UserWarning: Number of features (1) for attribute 'text_dense_features' does not match number of tokens (5). Set 'return_sequence' to true in the corresponding featurizer in order to make use of the features in 'CRFEntityExtractor'.` f"Number of features ({len(features)}) for attribute " Not sure how to fix that.

Which version of Rasa are you using? Could you please also share the content of your config.yml and (if you can) a minimal example of the NLU training data which is causing the bug?

Good day, I have a same problem. Version Rasa: 1.6.1. Training data and config.yml from example . In config.yml I replace language “en” with “en_core_web_md”.

Rasa Shell Issue

Rasa Open Source

Eli (Eli) January 29, 2020, 4:07pm 4

i am using Rasa 1.6.1

config.yml (313 Bytes)

I am not sure how to sure my NLU training data…

Topic		Replies	Views
UserWarning: Number of dense features (30) for attribute 'TEXT' does not match number of tokens (33) Rasa Open Source	1	289	May 14, 2021
ValueError: Sequence dimensions for sparse and dense features don't coincide Rasa Open Source	23	2178	February 11, 2020
NLU training data issue Rasa Open Source	2	481	October 16, 2020
large amount of memory. Rasa Open Source	4	1657	December 23, 2021
Rasa Training Killed unexpectedly Feedback on Rasa Open Source	5	915	October 3, 2022

Rasa Shell Issue

Related topics