Jeremy Macks

Still an issue for me even with Kaggle GPU's. Taking about three quarters of an epoch per hour. Was wondering if there was a checkpoint available or a way to...

As far as I know, this implementation is based off the papers, not actually identical to the paper itself, so we shouldn't expect identical accuracy rates.

> > > I see this in train.csv: > > DamslActTag,Text > "fo_o_fw_""_by_bc",Okay . > x, . *slash error > qh,Where to start. > sd,"I haven't had that much, of...

These are the guidelines written by the devs: https://github.com/JaidedAI/EasyOCR/blob/master/trainer/craft/README.md

https://github.com/clovaai/CRAFT-pytorch Due to intellectual property reasons, they will not release the training code :(. Unfortunately, there's no real solution to your problem, so hopefully you can find a workaround.

Are you using a custom model? I was unable to change hyperparameters programmatically on custom models.

Hmm, that's odd. Are you also calling blocklist? If you need a bandaid fix, you could potentially try using blocklist and restricting all non-numeric input. Annoying but would work.

Check to make sure you're not using different sized images as input. Since the model is fully convolutional, it tries to make spaces for as large of an input as...

Also having this problem, was curious if a solution was found. The free boxes get posted occasionally, but no consistently.

> I have a program using EasyOCR that works great, but when I try to change it to use EasyOCR's detection of text with rotation/skewness etc., as opposed to just...