fengxinjie/Transformer-OCR

Issues

• License

  #10 opened a year ago by yaysummeriscoming
  0

• 训练数据输入只能是96*96吗，一般的ocr识别模型32*200的大小好像不支持？

  #21 opened 3 years ago by apple2333cream
  0

• 请问这种报错是什么原因？

  #20 opened 3 years ago by lyy0215
  0

• Data format?

  #19 opened 4 years ago by sajedamurra
  2

• Transfer learning with new Vocab

  #18 opened 4 years ago by baoanh1310
  0

• Can't start the project

  #17 opened 4 years ago by apophatique
  0

• Any pretrained model?

  #1 opened 5 years ago by deep-practice
  8

• 训练数据输入问题

  #16 opened 4 years ago by yts19871111
  0

• Documentation on training and predicting

  #3 opened 4 years ago
  10

• what is the status of the paper?

  #14 opened 5 years ago by Jiakui
  0

• testloss would lead to model update on eval mode

  #7 opened 5 years ago by li10141110
  7

• About the learning rate

  #11 opened 5 years ago by leitro
  2

• auto config gpu/cpu

  #13 opened 5 years ago by eitamar-saraf
  0

• image is None

  #12 opened 5 years ago by eitamar-saraf
  0

• Why the recognition accuracy different from paper?

  #4 opened 5 years ago by zobeirraisi
  12

• UnicodeDecodeError: 'utf-8' codec can't decode byte 0xff in position 0: invalid start byte

  #8 opened 5 years ago
  12

• @delveintodetail Is there wrong in the predict.py file?

  #9 opened 5 years ago by delveintodetail
  2

• RuntimeError: cuda runtime error (59) : device-side assert triggered at /opt/conda/conda-bld/pytorch_1565272279342/work/aten/src/THC/generic/THCTensorScatterGather.cu:380

  #2 opened 5 years ago by wyc2015fq
  4

• Attention Visualization

  #6 opened 5 years ago by hiepph
  0

• You should carefully read the SVTP images.

  #5 opened 5 years ago by delveintodetail
  3