eragonruan/text-detection-ctpn issues and pull requests

#100 - In order to run this code using python3 make the following changes :-

Issue - State: open - Opened by arsalan993 over 7 years ago - 2 comments

#99 - I am getting this error while running it using python3. kindly help me resolve it

Issue - State: closed - Opened by arsalan993 over 7 years ago - 1 comment

#98 - prepare my own dataset

Issue - State: open - Opened by FantDing over 7 years ago - 1 comment

#97 - Where did you get your trainning dataset?

Issue - State: open - Opened by FantDing over 7 years ago - 3 comments

#96 - No overlaps when training

Issue - State: closed - Opened by cipri-tom over 7 years ago - 1 comment

#95 - Can I train the model based on the files in "checkpoints" dir

Issue - State: open - Opened by FantDing over 7 years ago - 5 comments

#94 - about "loss" when training the model

Issue - State: open - Opened by FantDing over 7 years ago - 7 comments

#93 - libcudart.so.9.1: cannot open shared object file: No such file or directory

Issue - State: closed - Opened by hsddlz over 7 years ago - 1 comment

#92 - I wonder while trainning new data where my new params saved?

Issue - State: open - Opened by qnkhuat over 7 years ago - 2 comments

#91 - training error( KeyError: 'max_overlaps')

Issue - State: closed - Opened by shuihuadx over 7 years ago - 1 comment

#90 - Is there faster rcnn/ssd/yolo based model for text detection?

Issue - State: closed - Opened by yt605155624 over 7 years ago - 1 comment

#89 - no attribute 'model_checkpoint_path'

Issue - State: open - Opened by FantDing over 7 years ago - 6 comments

#88 - Can ctpn detect English word instead of line?

Issue - State: closed - Opened by dajiangxiaoyan over 7 years ago

#87 - How to reduce false postive text lines?

Issue - State: open - Opened by dajiangxiaoyan over 7 years ago - 7 comments

#86 - 关于文字提取的训练

Issue - State: closed - Opened by KarsonHZ over 7 years ago

#85 - Any one successfully run the demo on linux + python 2.7 ?

Issue - State: open - Opened by keithbyr over 7 years ago - 3 comments

#84 - about detect speed

Issue - State: open - Opened by chengjunjiecn over 7 years ago - 4 comments

#83 - demo.py can't be run.

Issue - State: open - Opened by chengjunjiecn over 7 years ago - 1 comment

#82 - The implementation of vertical machanism

Issue - State: open - Opened by dodgaga over 7 years ago - 1 comment

#81 - ImportError: No module named torch

Issue - State: closed - Opened by rainman0709 over 7 years ago - 1 comment

#80 - Memory occupancy for a forward pass when launching demo.py

Issue - State: open - Opened by hellbago over 7 years ago - 1 comment

#79 - libstdc++.so.6: version `CXXABI_1.3.9' not found?

Issue - State: open - Opened by Haiyan-Chris-Wang over 7 years ago - 10 comments

#78 - different font size text detection setting

Issue - State: closed - Opened by wanghaisheng over 7 years ago - 7 comments

#77 - Detect individual words instead of whole line

Issue - State: open - Opened by guddulrk over 7 years ago - 7 comments

#76 - How can I get my own training data

Issue - State: open - Opened by LXYTSOS over 7 years ago - 5 comments

#75 - how do i prepare dataset

Issue - State: closed - Opened by sneffort over 7 years ago - 2 comments

#74 - how to get mlt_english+chinese dataset?

Issue - State: open - Opened by Xiangyaojun over 7 years ago - 4 comments

#73 - Run demo.py on windows with cpu successfully!

Issue - State: open - Opened by ZhuMingmin9123 over 7 years ago - 36 comments

#72 - change the number of anchors

Issue - State: open - Opened by cjt222 over 7 years ago - 4 comments

#71 - Fix bbox inside weights.

Pull Request - State: closed - Opened by walsvid over 7 years ago

#70 - Is there any difference between the reimplement of loss function and the formula in the paper?

Issue - State: closed - Opened by walsvid over 7 years ago - 5 comments

#69 - 关于训练效果的讨论

Issue - State: open - Opened by apple1987 over 7 years ago - 4 comments

#68 - About train dataset

Issue - State: closed - Opened by moondaiy over 7 years ago - 1 comment

#67 - Remove LSTM layer

Issue - State: open - Opened by JiteshPshah over 7 years ago - 1 comment

#66 - Detection not good for English words

Issue - State: open - Opened by arsalan993 over 7 years ago - 2 comments

#65 - text detection in vertical Images ((text flow in vertical direction) )

Issue - State: closed - Opened by JiteshPshah over 7 years ago - 5 comments

#64 - InternalError (see above for traceback): Blas GEMM launch failed

Issue - State: closed - Opened by kangshintae over 7 years ago - 1 comment

#63 - Detect text with different height

Issue - State: closed - Opened by guddulrk over 7 years ago - 3 comments

#62 - Invalid argument: ValueError: attempt to get argmax of an empty sequence

Issue - State: closed - Opened by bhanga-manish over 7 years ago - 1 comment

#61 - the training data batches just be 1 ?

Issue - State: closed - Opened by mymynew over 7 years ago - 1 comment

#60 - Train on our own dataset

Issue - State: closed - Opened by guddulrk over 7 years ago

#59 - ImportError: cannot import name 'bbox'

Issue - State: closed - Opened by GeneralJing over 7 years ago - 4 comments

#58 - Total Loss is so much fluctuating

Issue - State: closed - Opened by JiteshPshah over 7 years ago - 3 comments

#57 - saving intermediate checkpoint

Issue - State: closed - Opened by JiteshPshah over 7 years ago - 2 comments

#56 - bidirection lstm and fix some bug

Pull Request - State: closed - Opened by eragonruan over 7 years ago

#55 - why is lr so low?

Issue - State: closed - Opened by cjt222 over 7 years ago - 2 comments

#54 - fix bugs in lstm layer

Pull Request - State: closed - Opened by eragonruan over 7 years ago

#53 - Did you train your model on a Chinese database?

Issue - State: closed - Opened by houxiaoxia1 over 7 years ago - 5 comments

#52 - how to improve the performance in detecting small words?

Issue - State: closed - Opened by jycloud over 7 years ago - 23 comments

#51 - Thanks for writing this repo

Issue - State: closed - Opened by ghost over 7 years ago - 1 comment

#50 - SCALES and MAX_SIZE

Issue - State: closed - Opened by xrf116 over 7 years ago - 1 comment

#49 - loss become nan, invalid value encountered in log

Issue - State: closed - Opened by jycloud over 7 years ago - 5 comments

#48 - hello, i find ctpn has some trouble with small words!!!

Issue - State: closed - Opened by cjt222 over 7 years ago - 2 comments

#47 - oriented text connector

Pull Request - State: closed - Opened by eragonruan over 7 years ago

#46 - tensorboard

Issue - State: closed - Opened by zhao181 over 7 years ago - 4 comments

#45 - can not detect anything

Issue - State: closed - Opened by jsxyzb over 7 years ago - 2 comments

#44 - How To Setting Training Data Question

Issue - State: closed - Opened by Wangweilai1 over 7 years ago - 4 comments

#43 - Running with CPU only

Issue - State: open - Opened by zhao181 over 7 years ago - 68 comments

#42 - model can't converge

Issue - State: closed - Opened by hqliling over 7 years ago - 2 comments

#41 - how to deal with skew box

Issue - State: closed - Opened by cjt222 over 7 years ago - 1 comment

#40 - Problem with CPU model

Issue - State: closed - Opened by zhao181 over 7 years ago - 6 comments

#39 - about the detection speed

Issue - State: closed - Opened by apple1987 over 7 years ago - 1 comment

#38 - How to freeze the model graph from ckpt?

Issue - State: closed - Opened by wanghuok over 7 years ago - 6 comments

#37 - How can model run using multi gpus

Issue - State: closed - Opened by hsddlz over 7 years ago - 2 comments

#36 - Bus error: 10

Issue - State: closed - Opened by Paleve over 7 years ago - 5 comments

#35 - Isn't there a test.py to evaluate the model after training?

Issue - State: closed - Opened by xuewenyuan over 7 years ago - 1 comment

#34 - What is the format of gt file?

Issue - State: closed - Opened by ZhengWei0918 over 7 years ago - 1 comment

#33 - Bug In cpu only

Issue - State: closed - Opened by BaofengZan over 7 years ago - 1 comment

#32 - I really appreciate your work. And I saw the paper uses rnn(BLSTM). Did you use it in your code?

Issue - State: closed - Opened by zgy656 over 7 years ago - 1 comment

#31 - Would you kindly tell that only horizontal lines can be detected?

Issue - State: closed - Opened by lipond over 7 years ago - 2 comments

#30 - Now It is work! But I have a question about VGG16Net.

Issue - State: closed - Opened by rudebono over 7 years ago - 4 comments

#29 - exuse me !!!!do you have some tricks when training model?why do i train MLT dataset as default parameters and after 50000iters,it can detect nothing?

Issue - State: closed - Opened by cjt222 over 7 years ago - 15 comments

#28 - do you know what kind of datasets which ctpn uses? does the dataset built by themself and no release?

Issue - State: closed - Opened by cjt222 over 7 years ago - 1 comment

#27 - UserWarning: Converting sparse IndexedSlices to a dense Tensor of unknown shape. This may consume a large amount of memory. This may consume a large amount of memory.

Issue - State: closed - Opened by louisly over 7 years ago - 8 comments

#26 - ModuleNotFoundError: No module named 'lib'

Issue - State: closed - Opened by SuryaprakashNSM over 7 years ago - 12 comments

#25 - There are confusing parts.

Issue - State: closed - Opened by rudebono over 7 years ago - 4 comments

#24 - After CTPN. What is your idea?

Issue - State: closed - Opened by rudebono over 7 years ago - 10 comments

#23 - TypeError: 'NoneType' object is not subscriptable

Issue - State: closed - Opened by xiaomaxiao over 7 years ago - 4 comments

#22 - AttributeError: 'module' object has no attribute 'ndarray'

Issue - State: closed - Opened by louisly over 7 years ago - 1 comment

#21 - output of classfication is same !!!

Issue - State: closed - Opened by cjt222 over 7 years ago - 3 comments

#20 - Bug in cpu only mode, maximum recursion depth exceeded in comparison

Issue - State: closed - Opened by xrf116 over 7 years ago - 3 comments

#19 - Problem

Issue - State: closed - Opened by rudebono over 7 years ago - 1 comment

#18 - What files should be in the gt_path folder?

Issue - State: closed - Opened by rudebono over 7 years ago - 3 comments

#17 - Is there a way to train my own image data?

Issue - State: closed - Opened by rudebono over 7 years ago

#16 - Could you put your dataset and model in baidu yunpan?

Issue - State: closed - Opened by chinakook over 7 years ago - 2 comments

#15 - cuda nms + loss function

Pull Request - State: closed - Opened by eragonruan over 7 years ago

#14 - rpn_loss_box is always equal 0

Issue - State: closed - Opened by cjt222 over 7 years ago - 6 comments

#13 - fatal error: numpy/arrayobject.h: No such file or directory

Issue - State: closed - Opened by xuewenyuan over 7 years ago - 2 comments

#12 - BiLSTM and Training Time

Issue - State: closed - Opened by TaoDream almost 8 years ago - 14 comments

#11 - kernel not found in checkpoint

Issue - State: closed - Opened by hsddlz almost 8 years ago - 1 comment

#10 - Create LICENSE

Pull Request - State: closed - Opened by eragonruan almost 8 years ago

#9 - python2/3 compatiblity

Pull Request - State: closed - Opened by eragonruan almost 8 years ago

#8 - RuntimeError: Graph is finalized and cannot be modified.

Issue - State: closed - Opened by BrownOfSummer almost 8 years ago - 1 comment

#7 - I try to change code from python2 to python3

Issue - State: closed - Opened by cjt222 almost 8 years ago - 14 comments

#6 - split_label.py deal with .txt but .xml provided in VOC2007, and import python_nms error in demo

Issue - State: closed - Opened by BrownOfSummer almost 8 years ago - 6 comments

GitHub / eragonruan/text-detection-ctpn issues and pull requests