Roee Litman

Hi Thanks for your interest in this package. As far as I recall we had, on some of the datasets, a token/character for white-space. you can use it to generate...

Hi @kurapan . I am afraid that the exact conditions for reproducing these experiments are a little hard for us to recall, as we did not keep the exact hyper-parameters...

Hi We released the IAM weights [here](https://github.com/amzn/convolutional-handwriting-gan/releases/tag/CVPR20). AFAIK (or remember) testing code is a part of the clovaa OCR repo we mention in the paper.

You'd have to forgive me, but it has been a long time since we looked at this code... Please have a look in the [LMDB generation code](https://github.com/amzn/convolutional-handwriting-gan/blob/master/generate_wordsLMDB.py), it iterates over...

Did you try to run the regular IAM/RIMES experiment from scratch?

Hi @kymillev and @AhmedAl93, Thanks for you interest in this package, and for you patience. We try our best to respond to your questions in this challenging times. The common...

We so not plan to release pre-trained model on RIMES, sorry.

Hi @rlgnswk, thanks! You can look at it as "pre-training", we used the 3rd **row** as an initialization for the "regular" supervised training. The ± is std over a few...

I hope you do not mean the GAN recognizer, as we intentionally made it sub-optimal (see details in the paper and the supplementary materials). We use the [clova text recognition...

Please make sure you're making correct+consistent use of character encoding (should be UTF-8 by default in py3, no?) throughout `*.py` files as well as other file types as used. We...