AdvancedLiterateMachinery

AdvancedLiterateMachinery copied to clipboard

I'm working with google colab & try to build the DCNv2 model to train LORE-TSR on my custom dataset. when trying to run the: !bash make.sh command i always encounter...

ellamiamaria

vgt能否识别中文版面？该如何实现呢？

zhuxiaobin

Refactor VIEDataset by removing unused `self.dataset`

1

Utilize `self.dataset_root_path`

Isydmr

I tried finetuning the wireless model on my private dataset but during the inference i could see that the bbox of all the cells were shifted towards the top right:...

vavinash992

有大神写一个吗？谢谢！

dyh

How to make a data set in Funsd format?

HHHEDA

I wonder if there is any script that can help to reconstruct (concatinating `text_lists` w.r.t to the polygon positions) in order to form text blocks ? is there any such...

naarkhoo

Dictionary comprehension is typically faster and more Pythonic than using a for loop.

naarkhoo

there are cases where the text_recognizer fail due to resize from modelscope

naarkhoo

Update text_detection.py

1

if in a whole document (e.g. pdf file) there is one page with picture and no text your existing pipeline fails because of empty "detection". so I am just returning...

naarkhoo