clovaai/deep-text-recognition-benchmark issues and pull requests

#99 - Sample ipython notebook demonstrating use of the code

Pull Request - State: closed - Opened by tgalkovskyi almost 5 years ago - 1 comment

#98 - Transfer learning with pretrained model on custom dataset

Issue - State: closed - Opened by aashish-0393 almost 5 years ago - 4 comments

#97 - Add SPACE to target character sets

Issue - State: closed - Opened by thaihaquang almost 5 years ago - 2 comments

#96 - changing model parameters with pretrained weights

Issue - State: closed - Opened by mharsh1903 almost 5 years ago - 4 comments

#95 - removed align_corners=True

Pull Request - State: closed - Opened by varshaneya almost 5 years ago - 1 comment

#94 - Training iter error

Issue - State: closed - Opened by DLH06 almost 5 years ago - 1 comment

#93 - How to set the proper parameter under the different scene?

Issue - State: closed - Opened by 13438960761 almost 5 years ago - 2 comments

#92 - Use --Trans TPS --Pad and set ImgW = 640

Issue - State: closed - Opened by liangzimei almost 5 years ago - 1 comment

#91 - updated grid_sample arg align_corners=True

Pull Request - State: closed - Opened by edwardpwtsoi almost 5 years ago - 1 comment

#90 - about Word vector

Issue - State: closed - Opened by WenmuZhou almost 5 years ago - 1 comment

#89 - Training time

Issue - State: closed - Opened by SealQ almost 5 years ago - 1 comment

#88 - Will the results of FAN be released?

Issue - State: closed - Opened by liuzhuang1024 almost 5 years ago - 2 comments

#87 - Could you kindly provide the premade lmdb 'data_lmdb_release.zip' md5 code?

Issue - State: closed - Opened by sonack almost 5 years ago - 2 comments

#86 - Could I get some help for want to make it work in Korean?

Issue - State: closed - Opened by FKgk almost 5 years ago - 2 comments

#85 - To train on my own dataset

Issue - State: closed - Opened by xxxpsyduck almost 5 years ago - 30 comments

#84 - Is the pretrained model trained on english only?

Issue - State: closed - Opened by xxxpsyduck almost 5 years ago - 1 comment

#83 - bug in the code: do not need to ignore index for attention loss

Issue - State: closed - Opened by linhuifj almost 5 years ago - 2 comments

#82 - how to get confidence score of each word using Attn

Issue - State: closed - Opened by nbhupendra almost 5 years ago - 2 comments

#81 - https://github.com/clovaai/deep-text-recognition-benchmark/issues/78#issuecomment-535885352

Issue - State: closed - Opened by nbhupendra almost 5 years ago - 1 comment

#80 - a question

Issue - State: closed - Opened by softwaregitgub almost 5 years ago - 1 comment

#79 - Why I don't need to train TSP network seperately .

Issue - State: open - Opened by liquor233 almost 5 years ago

#78 - how to get accuracy score of each word

Issue - State: closed - Opened by nbhupendra almost 5 years ago - 6 comments

#77 - --rgb

Issue - State: closed - Opened by HJiangXin almost 5 years ago - 1 comment

#76 - [Convert to onnx]

Issue - State: open - Opened by trungna1374 almost 5 years ago - 17 comments

#75 - .pth file how to used

Issue - State: closed - Opened by HJiangXin almost 5 years ago - 2 comments

#74 - how to handle corrupt imgs in lmdb dataset

Issue - State: closed - Opened by crazysal almost 5 years ago - 1 comment

#73 - not able to change model parameters

Issue - State: closed - Opened by mharsh1903 about 5 years ago - 4 comments

#72 - How to detect special characters?

Issue - State: closed - Opened by kalai2033 about 5 years ago - 1 comment

#71 - predict error

Issue - State: open - Opened by 13438960761 about 5 years ago

#70 - image size

Issue - State: closed - Opened by mharsh1903 about 5 years ago - 3 comments

#69 - Gap between the accuracy of paper reported and that of released model

Issue - State: closed - Opened by jeffreykuang about 5 years ago - 1 comment

#68 - different decode operation between ctc and attention in converter

Issue - State: closed - Opened by bityigoss about 5 years ago - 1 comment

#67 - How to Use Multiple GPUs.

Issue - State: closed - Opened by inouetaka about 5 years ago - 2 comments

#66 - CTCloss gets the nan loss when training with a custom Chinese dataset.

Issue - State: open - Opened by AnddyWang about 5 years ago - 20 comments

#65 - path does not exist error while creating custom dataset

Issue - State: closed - Opened by omersert about 5 years ago - 6 comments

#64 - How do you generate the SynthText dataset?

Issue - State: closed - Opened by eepaul about 5 years ago - 2 comments

#63 - unable to recognise text with numbers

Issue - State: open - Opened by p9anand about 5 years ago - 2 comments

#62 - what is the difference between data_lmdb_baidu.xx and data_lmdb_reuploaded.xx?

Issue - State: closed - Opened by eepaul about 5 years ago - 2 comments

#61 - Multilabel labels in one image

Issue - State: open - Opened by anmol4210 about 5 years ago - 3 comments

#60 - Why the Accuracy of training CTC model is equal=0.0 after 50000 Iter.?

Issue - State: closed - Opened by zobeirraisi about 5 years ago - 1 comment

#59 - cannot download data

Issue - State: closed - Opened by CuriousCat-7 about 5 years ago - 2 comments

#58 - how to output the probability of atten?

Issue - State: closed - Opened by Juandai8401 about 5 years ago - 4 comments

#57 - why error occured create_lmdb_dataset?

Issue - State: closed - Opened by parksunwoo about 5 years ago - 3 comments

#56 - Accuracy difference between local retraining model and pretrained one

Issue - State: closed - Opened by 1LOVESJohnny about 5 years ago - 11 comments

#55 - set_storage is not allowed on Tensor created from .data or .detach()

Issue - State: closed - Opened by Crescentz about 5 years ago - 5 comments

#54 - Does pretrained model support korean?

Issue - State: closed - Opened by kboseong about 5 years ago - 1 comment

#53 - error of training

Issue - State: closed - Opened by xkyi about 5 years ago - 5 comments

#52 - when i running train.py,have an error ValueError: num_samples should be a positive integer value, but got num_samples=0

Issue - State: closed - Opened by HJiangXin about 5 years ago - 5 comments

#51 - best accuracy: 0.000 after train 200000 steps

Issue - State: closed - Opened by tienduchoang about 5 years ago - 4 comments

#50 - fixed image size

Issue - State: closed - Opened by chenjun2hao about 5 years ago - 3 comments

#49 - Can't training model with own lmdb dataset

Issue - State: closed - Opened by thangtran480 about 5 years ago - 8 comments

#48 - replaced deprecated fromstring() with frombuffer()

Pull Request - State: closed - Opened by YacobBY about 5 years ago - 1 comment

#47 - requires_grad

Issue - State: closed - Opened by soldierofhell about 5 years ago - 1 comment

#46 - Update README.md

Pull Request - State: closed - Opened by coallaoh about 5 years ago

#45 - probelm about attention

Issue - State: closed - Opened by xiaoshuizhu202 about 5 years ago - 1 comment

#44 - Add checks for cuda to make software run without GPU

Pull Request - State: closed - Opened by jotsif about 5 years ago - 4 comments

#43 - recog error using TPS-ResNet、VGG-BiLSTM-Attn

Issue - State: closed - Opened by AnddyWang about 5 years ago - 16 comments

#42 - Conversion to ONNX

Issue - State: open - Opened by Sayyam-Jain about 5 years ago - 8 comments

#41 - test ctc model it is giving runtime error?

Issue - State: closed - Opened by mythuven about 5 years ago - 3 comments

#40 - Train --sensitive mode with the MJSynth + ST_spe

Issue - State: closed - Opened by mahendra047 about 5 years ago - 2 comments

#39 - Simple image causing troubles

Issue - State: closed - Opened by marcoromelli about 5 years ago - 6 comments

#38 - The text shape when using Attention

Issue - State: closed - Opened by SakuraRiven about 5 years ago - 1 comment

#37 - Does the model support variable length recognition？I train the model using Chinese STR Datasets，but have a bad accuracy

Issue - State: closed - Opened by AnddyWang about 5 years ago - 4 comments

#36 - Nvidia Apex for FP16 calculations

Pull Request - State: closed - Opened by YacobBY about 5 years ago - 3 comments

#35 - Results different with online Demo

Issue - State: closed - Opened by crazysal about 5 years ago - 1 comment

#34 - use the demo script，GPU OOM

Issue - State: closed - Opened by AnddyWang about 5 years ago - 1 comment

#33 - Can the model be used for Chinese STR case retrained?

Issue - State: closed - Opened by iPrayerr about 5 years ago - 2 comments

#32 - strange accuracy in IC03

Issue - State: closed - Opened by techkang about 5 years ago - 1 comment

#31 - Does ctc requires a blank character ?

Issue - State: closed - Opened by thomasryck about 5 years ago - 2 comments

#30 - Thoughts about generated/real images ratio

Issue - State: closed - Opened by thomasryck about 5 years ago - 1 comment

#29 - Time it takes to train ResNet + CTC and VGG + CTC

Issue - State: closed - Opened by alwc about 5 years ago - 5 comments

#28 - No bn layers in CRNN ?

Issue - State: closed - Opened by jianxiangm about 5 years ago - 1 comment

#27 - Access to other models?

Issue - State: closed - Opened by QuickLearner171998 about 5 years ago - 3 comments

#26 - Difference in performance between online demo website and the offline code

Issue - State: closed - Opened by AyushP123 about 5 years ago - 7 comments

#25 - Does not work if CUDA v10 is installed.

Issue - State: closed - Opened by QuickLearner171998 about 5 years ago - 2 comments

#24 - using this model for a scene with multiple lines of text

Issue - State: closed - Opened by MattHuntebrinker about 5 years ago - 2 comments

#23 - The loss converge but the accuracy always stay zero with optimizer adam.

Issue - State: closed - Opened by tinggh about 5 years ago - 2 comments

#22 - quadrangle / rotated bbox

Issue - State: closed - Opened by soldierofhell about 5 years ago - 3 comments

#21 - vertical text recognition

Issue - State: closed - Opened by mathpopo about 5 years ago - 2 comments

#20 - Unexpected key(s) in state_dict

Issue - State: closed - Opened by MattHuntebrinker about 5 years ago - 2 comments

#19 - training is ZeroDivisionError:division by zero and how to automatically generate gt.txt file

Issue - State: closed - Opened by Changs92 about 5 years ago - 1 comment

#18 - Can't seem to recognize numbers

Issue - State: closed - Opened by ziudeso about 5 years ago - 1 comment

#17 - TypeError: can't pickle Environment objects

Issue - State: closed - Opened by MattHuntebrinker about 5 years ago - 2 comments

#16 - miss match in size

Issue - State: closed - Opened by GSATHYANARAYANA about 5 years ago - 7 comments

#15 - Do you have any plans to release the pretrained model on ICDAR2019 ReCTS (task1)?

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

#14 - How many iterations or loss, can I have a so-so results on my test images ?

Issue - State: closed - Opened by XiaoLaoDi over 5 years ago - 8 comments

#13 - different accuracy between paper and competition website

Issue - State: open - Opened by bliu3650 over 5 years ago - 19 comments

#12 - fix: 중복된 코드 삭제

Pull Request - State: closed - Opened by seominseok0429 over 5 years ago

#11 - Inference code

Issue - State: closed - Opened by ghost over 5 years ago - 9 comments

#10 - I cannot download the pretrained model in google drive. Is there a resource in baidu Netdisk?

Issue - State: closed - Opened by tinggh over 5 years ago - 13 comments

#9 - Image's width should be fixed?

Issue - State: closed - Opened by 1033020837 over 5 years ago - 2 comments

#8 - Update test.py

Pull Request - State: closed - Opened by Sangkwun over 5 years ago - 1 comment

#7 - the best models case insensitive?

Issue - State: closed - Opened by xiaoyubing over 5 years ago - 11 comments

#6 - new question

Issue - State: closed - Opened by CXY573 over 5 years ago - 4 comments

#5 - Unpack error.

Pull Request - State: closed - Opened by Sangkwun over 5 years ago - 1 comment

GitHub / clovaai/deep-text-recognition-benchmark issues and pull requests