search

tyiannak / pyAudioAnalysis

Python Audio Analysis Library: Feature Extraction, Classification, Segmentation and Applications
Apache License 2.0
5.83k stars 1.19k forks source link

Example output not matching 2022 output #365

Open qkum opened 2 years ago

qkum commented 2 years ago

Copy pasted the code and even downgraded scikit to the correct old version.

Wiki : 2022 2022 output: 2022_2

Pip list: 2022_3

Python --version: 2022_4

qkum commented 2 years ago

Found one more. This time with python 3.7:

1

2022, maybe the model code changed?: 2

qkum commented 2 years ago

If you can't make the test results match the wiki/documentation/readme results the respiratory is pretty useless sadly.

Someone who can verify what is correct and wrong should set up a docker container or images for the future.

This is as bad as it comes unless you don't care if your results are correct or not 😞

tyiannak commented 2 years ago

wrt the aS.mid_term_file_classification() example: the wiki was correct, it was just missing the aggregated statistics printed in command line and the overall accuracy. I've just added it.

wrt the speaker diarization problem: there's indeed difference probably due to some of the dependencies that has changed. I'll try to figure this out and will update in the next days.

And yes, I know there should be some pytests for all 4 basic functionalities (feature extraction, classification, regression and segmentation), I just don't had the time to add this yet.

tyiannak commented 2 years ago

Diarization should be ok now (check master). Also, the Wiki is updated with the latest diarization output example https://github.com/tyiannak/pyAudioAnalysis/wiki/5.-Segmentation