Replicating papers with pytorch

Whether the code implemented by tensorflow can be multiplexed with pytorch. Recently, many papers have been reproduced using tensorflow, but I want to change to pytorch, such as CTPN, attention-ocr, etc.

There should be no reason, why these papers can’t be implemented in PyTorch.
If you encounter any issues, e.g. if some operations seem to be missing, please post them in this forum.

1 Like

alright, thank you very much!:grinning: