bgshih/crnn

Issues

• after getting the lmdb dataset, where should I put them to ?

  #106 opened 7 years ago
  1

• impossible to run it on ubuntu 16.04?

  #105 opened 7 years ago
  6

• impossible to run it on ubuntu 17.04?

  #104 opened 7 years ago
  0

• sh build_cpp.sh

  #103 opened 7 years ago
  10

• THPP_LIBRARY AGAIN !!

  #102 opened 7 years ago
  5

• manifest for kaixhin/cuda-torch:latest not found

  #101 opened 7 years ago
  3

• On which dataset it is currently trained on?

  #100 opened 7 years ago
  1

• About the irregular scene text dataset?

  #99 opened 7 years ago
  4

• can you or anyone upload all the dependence (folly, thpp fbthrift fblualib)zip file ?

  #98 opened 7 years ago
  5

• Failed to create an image using the supplied Dockerfile

  #97 opened 7 years ago
  5

• Use trained model for CPU only machine?

  #96 opened 7 years ago
  0

• C or CPP interface demo.lua

  #95 opened 7 years ago
  0

• CRNN Prediction is Very Slow

  #94 opened 7 years ago
  0

• Is there a special training strategy(include pre-processing)?

  #93 opened 7 years ago
  1

• converge problem

  #92 opened 7 years ago
  0

• requirement

  #91 opened 7 years ago
  0

• Hardware Environment?

  #90 opened 7 years ago
  1

• tensorflow

  #89 opened 7 years ago
  0

• Meaning of parameter？

  #88 opened 7 years ago
  4

• how to use tensorflow to instead of torch ?

  #87 opened 7 years ago
  2

• Imbalanced classes

  #86 opened 7 years ago
  2

• @ bgshih i not install torch , Can i use pytorch ? i use pytorch,when i sh build_cpp.sh,there has many errors, is it must use torch,not pytorch?

  #85 opened 7 years ago
  2

• Stuck with the installation of the Dependencies. Is any One Got a Stable Version of these Libraries

  #84 opened 7 years ago
  0

• is CPU option available ?

  #83 opened 7 years ago
  3

• Training fails, Tensor.lua:462: bad argument #1 to 'set'

  #82 opened 7 years ago
  4

• Paper is 25 (25 frames) but code is 26 (26x512).

  #81 opened 6 years ago
  3

• How to train crnn using pictures with different width.

  #80 opened 7 years ago
  1

• training no convergence, why?

  #79 opened 7 years ago
  1

• Changing the maximum length of string

  #78 opened 7 years ago
  2

• How CTC deals with successive doubled characters ?

  #77 opened 7 years ago
  5

• How blank/ null class in CTC allows to reduce memory burden on the network and allows repeated labels ?

  #76 opened 7 years ago
  0

• State-of-the-art

  #75 opened 7 years ago
  1

• Why preds variable contains negative values?

  #74 opened 7 years ago
  1

• embedding layer

  #73 opened 7 years ago
  0

• Installed torch with LUA51 - failed in Assertion `lua_gettop(L) == top'

  #72 opened 7 years ago
  2

• What's dataset the pretrained model use?

  #71 opened 7 years ago
  2

• a question regard with testing

  #70 opened 7 years ago
  0

• libcrnn not found

  #69 opened 8 years ago
  4

• fb

  #68 opened 8 years ago
  0

• Work on line level instead of word level

  #67 opened 6 years ago
  20

• How can one explain that CRNN comes from VGG deep net ?

  #66 opened 7 years ago
  0

• Learning in crnn

  #65 opened 7 years ago
  0

• Error differential backpropagation in CRNN

  #64 opened 7 years ago
  0

• Preprocessing of Training Images

  #63 opened 8 years ago
  3

• Quick question about crnn_demo_model.t7

  #62 opened 8 years ago
  2

• Training on Different Dataset

  #61 opened 8 years ago
  10

• Recognition of Special Characters

  #60 opened 8 years ago
  1

• Using UTF-8 Character encoding instead of ASCII

  #59 opened 8 years ago
  1

• CNN

  #58 opened 7 years ago
  0

• How do you explain that network predicts blank firsr “ --hh-e-l-ll-oo-- ”?

  #57 opened 7 years ago
  2