Matteo

@osayes thanks for your response. I think you probably got the point, but if this is the reason, shouldn't it be a better behaviour from tensorflow library to execute everything...

I verified I'm running CUDA version 11.2, cuDNN 8.1 and python 3.9 from conda environment, so Iìm running in the first tested configuration running `tf.device("/gpu:0")` always returns this message: ```...

as expected cupy uses only my GPU for this operation! It uses 100% of the graphic card without touching the CPU. Thank you @osayes I didn't know this library :)

UPDATE: Don't know why, but just __after updating PopOS seems to work properly__ (I don't know if only this or something in the last hundreds of commits fixed my issue...)....

Thank you for your kind support.

Just made a new PR (#5233) for index page, which result still untranslated, based on PR #1576 by @csr.

Welcome @baleDevIt :) it seems like there isn't one, as we are just few people I think anyone can see others translations and if he likes can assign himself a...

Since it was a one year old PR I updated the translation with @Ale-Cas contributions and I pulled the latest code update from master branch.

Is there, at the moment, any way to color the output of the table?