Noam Bressler
Noam Bressler
About the two points: 1. I think only now I fully understood what you meant. I agree with you that I'd prefer a metric what would return 1 also for...
Start after https://github.com/deepchecks/deepchecks/issues/1900 is done. ## Tasks: * Add a flag that instead of computing tf-idf, the check computes the frequency of the word in the dataset as a whole...
Should be considered together with https://github.com/deepchecks/deepchecks/issues/1928
> Related to #1394 Would even say duplicate? If so @TheSolY Maybe copy the stuff from there to here and close #1394
Reopened as #2128 does not measure blurriness, rather "texture smoothness" or something like it
Thanks for raising the issue. We will try to resolve this in one of the next releases.
## Tasks * Create code similar to get_feature_label_correlation_per_class but without the change part of the calculation and display. Build it so it'll apply also for tabular features. Refactor so that...
Currently - the only thing is adding the NLP Check section
Addressed many comments. Stuff that is not implemented yet: 1. Token classification format still doesn't work with Dataset object 2. Documentation for formats in docstrings 3. tests 4. Calculating metrics
What's new: SingleDatasetPerformance now works for text classification, including multi label What's missing: 1. Tests 2. Docs 3. Token classification (need to solve seqeval dependency issue) 4. Not fully tested...