ocr-d-modul-2-segmentierung / page-segmentation

Pixel classifier for historic prints. Mirror of https://gitlab2.informatik.uni-wuerzburg.de/ls6/ocr4all-pixel-classifier. Report issues here.
Other
5 stars 2 forks source link

Tensorflow requirements #8

Open stweil opened 2 years ago

stweil commented 2 years ago

Is the current limitation which only allows tensorflow 2.5.0 really necessary? Why not 2.5.1, 2.6 or even the very latest tensorflow?

Such old versions of Tensorflow are for example not provided for Python 3.10 (which only supports tensorflow 2.8.0), and even for Python 3.7 and other Python versions it causes version conflicts with other packages which are required for OCR-D.

crater2150 commented 2 years ago

As explained in #5, having dynamic dependencies broke the installation recently, and I don't have time to regularly check versions of all transitive dependencies. If you can test installation with newer versions, I'll accept PRs updating the requirements file.