AstarLight/Lets_OCR

Issues

• 这里 coords_left 和 coords_right 并无法保证相等

  #101 opened 3 months ago by kingzcheung
  1

• 遇到的一些问题和疑惑（creat_lmdb_dataset 和convert模块）

  #102 opened 5 months ago by fyy0322
  0

• ctpn跑了两天 终于跑通代码了，太难了，把跑代码过程遇到问题都给整合，希望能帮到大家

  #94 opened 4 years ago by CHENG-EMMA1
  15

• 关于文本中的空格如何识别

  #64 opened 5 years ago by wting861006
  1

• 训练时Corrupted image errror，字节对象进lmdb，lmdb读出来的不一样，请高手帮解决，急。

  #75 opened 5 years ago by luyun760324
  7

• 提供的云盘数据跟代码里的alphabets根本对不上号。。。

  #90 opened 4 years ago by rsqai
  1

• 提供标签（图像路径 中文字符）

  #82 opened 2 years ago by zchoi
  4

• infer w160_bs64_model下的模型，识别结果都是错误的

  #93 opened 2 years ago by hwangyl
  1

• 使用自己的训练集从头训练crnn，accuracy一直为0

  #100 opened 3 years ago by iakRulan
  0

• need help for training with custom images

  #99 opened 3 years ago by anidiatm41
  0

• 训练问题

  #87 opened 5 years ago by Waynepoo
  2

• 关于ctpn与crnn整合问题

  #98 opened 3 years ago by InferMaster
  1

• 可以帮忙提供数据集吗

  #97 opened 4 years ago by 2291754463
  0

• 学习代码中的一些问题

  #96 opened 4 years ago by yangshiju
  0

• 请问可否使用这个代码作为毕设用途

  #95 opened 4 years ago by liangtao123
  0

• warning: img ./train_data/data_ready/MSRA_TD500/train_im/img_1756.jpg raise error!

  #76 opened 5 years ago by stephonjou
  6

• Got a error of " can not import name 'gen_geo_map' when running train.py for east part

  #92 opened 4 years ago by lmw0320
  1

• 把训练数据转换lmdb之前，标签是文字还是对应的数字？？？

  #89 opened 4 years ago by jiaoxiaosong
  5

• 付费向学霸哥哥请教

  #70 opened 5 years ago by wangzazao
  2

• 运行 infer.py 时提示 No module named 'lib.nms'

  #67 opened 5 years ago by hanhongchang
  3

• 在docker里sh test.sh出错No module named nms

  #78 opened 5 years ago by Law0825
  1

• module 'lib.nms' has no attribute 'cpu_nms'

  #85 opened 5 years ago by Lie-huo
  1

• 这模型实现是错误的吧

  #91 opened 4 years ago by yuxx0218
  1

• MaxPool2d中窗口大小与步长的问题

  #88 opened 5 years ago by jiangmeixian
  0

• 运行train.py出现Corrupted image for 5372问题，求各位大神指教

  #73 opened 5 years ago by ShangLe0607
  2

• 有谁能解决Corrupted image fo 我看很多都遇到这个错

  #81 opened 5 years ago by paohaijiao
  1

• 关于crnn训练的自定义数据集准备

  #86 opened 5 years ago by wting861006
  0

• 图片生成

  #84 opened 5 years ago by zcswdt
  0

• Corrupted image errror

  #61 opened 5 years ago by ahbon123
  5

• 准确率与召回率

  #83 opened 5 years ago by PKU-FengShuo
  0

• 2 question about your CTPN code

  #80 opened 5 years ago by tonykcn
  0

• docker 是不是不能用了？

  #79 opened 5 years ago by Bigmai-1234
  1

• draw_image模块问题

  #63 opened 5 years ago by Azhihong
  2

• 训练报错 line 90, in backward allow_unreachable=True) # allow_unreachable flag RuntimeError: CUDA error: an illegal memory access was encountered，求解决问题

  #77 opened 5 years ago by luyun760324
  0

• 我用自己的训练集转lmdb，CRNN训练时loss都是20多，不收敛。我只识别数字和字母

  #74 opened 5 years ago by ZhangZoffy
  7

• train 数据集时候，未收敛便提前终止了，并且和题主给的训练图片也不太一样，len(train_loader)输出的是7，不知道哪里出错了

  #68 opened 5 years ago by joeqincai
  3

• train.py 中TypeError: can't pickle Environment objects错误？

  #66 opened 5 years ago by joeqincai
  3

• 编码问题，求解答

  #71 opened 5 years ago by ShangLe0607
  0

• CTC计算出来的loss非常大，大于60的loss, 不知道哪里出了问题

  #72 opened 5 years ago by joeqincai
  10

• train 数据集时候，未收敛便提前终止了，并且和题主给的训练图片也不太一样，len(train_loader)输出的是7，不知道哪里出错了

  #69 opened 5 years ago by joeqincai
  0

• 请问长文本图片是先切好再做inference再考虑整合吗？

  #65 opened 5 years ago by zzks
  0

• 识别问题

  #59 opened 5 years ago by Crescentz
  1

• ctpn项目中，为什么所有anchor的bottom坐标都比top坐标要大

  #60 opened 5 years ago by Nobody0321
  1

• 关于ctpn的源码的一些问题集中提问

  #62 opened 5 years ago by Nobody0321
  1

• 训练数据

  #58 opened 5 years ago by ziqi3
  0

• RuntimeError: set_sizes_contiguous is not allowed on Tensor created from .data or .detach()

  #57 opened 5 years ago by wjx-git
  1

• AssertionError: index range error

  #56 opened 5 years ago by wjx-git
  1

• 训练自己的数据集

  #55 opened 5 years ago by seyoulala
  1

• 运行infar检测一张图像过程CPU会爆掉

  #54 opened 5 years ago by GanWang
  0

• 请问文本识别的数据集有数据集的描述吗？比如数据集用了哪些字体之类的

  #53 opened 5 years ago by vincyqian
  0