vugia truong

I make some changes for japanese https://github.com/gachiemchiep/SynthText i should work with chinese too. You just need to replace the font and run again. Hope this help someone who interest

@ZLeopard So did you mean that currently "Free Image Size" feature does not work as the description ???

@ZLeopard I will go deeper into the source and look for further information

@barryluke96 Some thing like this Image of the mask  with timer.env('Postprocess'): save = cfg.rescore_bbox cfg.rescore_bbox = True classes, scores, boxes, masks = postprocess(dets_out, w, h, visualize_lincomb = args.display_lincomb, crop_masks...

Anyone solved this problems ? I tried everything suggested here such as : reduce lr, turning off fp16, reduce batch size, turn off cache. I even switch yolox back to...

@Tailor2019 Sorry I don't understand your question.

@Tailor2019 Each image is generated using 1. randomly selected text from a dictionary file 2. randomly selected font from font directory 3. random select color value from a color range...

@Tailor2019 The logic of using background images is defined inside this method. https://github.com/Belval/TextRecognitionDataGenerator/blob/ab83b94fd10ecdace77c77fddb2727d8e4c85289/trdg/background_generator.py#L58 The background image is randomly crop and resize. You can change it to fit your need.

@AICampB4 @aziz-ullah-khan That's odd. I test on my machine and it works well. ```bash (py3.9) gachiemchiep:tmp$ trdg -l ja -c 10 -w 5 (py3.9) gachiemchiep:tmp$ pip freeze | grep trdg...

Hi @mdbloice . Augmentor is great tool. It did help my work a lot. Hope you release the new update soon. Best regard