lexpredict-lexnlp icon indicating copy to clipboard operation
lexpredict-lexnlp copied to clipboard

Published 20 hours ago • verified publisher icon LexPredict

ValueError - sections.get_sections()

Open dmnxprss opened this issue 6 years ago • 7 comments

Encountering a ValueError when running get_sections()

Specifically:

Number of features of the model must match the input

Number Features in Test: 313 Features in Model: 369

It seems not all the same features are being captured within build_section_break_features()?

dmnxprss avatar Aug 17 '18 18:08 dmnxprss

@dmnxprss , thanks for reporting this. Few quick questions:

  1. Can you share the version you're using for lexnlp?
  2. Have you customized any models?
  3. Have you used virtualenv/conda and/or a global python installation to install?

mjbommar avatar Aug 18 '18 16:08 mjbommar

  1. V 2.0 (Aug Release)
  2. No customization of any models
  3. Global Python 3.6 install

dmnxprss avatar Aug 21 '18 13:08 dmnxprss

@dmnxprss , had you previously installed an earlier version of LexNLP?

mjbommar avatar Aug 21 '18 13:08 mjbommar

Yes, I believe I previously had 1.8 installed

dmnxprss avatar Aug 23 '18 23:08 dmnxprss

I am encountering precisely the same error: v0.2.3 No customization Conda environment - Python 3.6.6

~~Possibly also of relevance is the following warning: "UserWarning: Trying to unpickle estimator DecisionTreeClassifier from version 0.19.0 when using version 0.19.1. This might lead to breaking code or invalid results. Use at your own risk."~~

phi2039 avatar Jan 29 '19 03:01 phi2039

Hi All, Checking in to see if there has been a resolution to this bug?

Thank You

dmnxprss avatar Mar 18 '19 16:03 dmnxprss

Hey Guys,

Just following up here? Is there a new section_segmenter.pickle file we could use which has been trained on 313 features to allow us to use get_sections() function?

Thanks

dmnxprss avatar Aug 12 '19 19:08 dmnxprss