Minimal port to Tensorflow 2

emedvedev / attention-ocr

A Tensorflow model for text recognition (CNN + seq2seq with visual attention) available as a Python package and compatible with Google Cloud ML Engine.

MIT License

1.08k stars 256 forks source link

Minimal port to Tensorflow 2 #198

Closed mjpieters closed 1 year ago

mjpieters commented 1 year ago

The changes in this PR are the absolute minimum needed to get this project to work with Tensorflow 2.

I am a ML and TF novice, so I probably missed something, but these changes allowed me to use attention-ocr to train a new model and produce predictions for new inputs, using only TF2.

emedvedev commented 1 year ago

That's awesome work, thank you for this. Merging :)