baudm/parseq issues and pull requests

#156 - stops at Validation sanity check

Issue - State: open - Opened by fuzheng1209 10 days ago

#155 - After converting the PyTorch model to ONNX with dynamic batch size, if the input image quality is low, the eos_id cannot be retrieved.

Issue - State: open - Opened by Whoo-jl 14 days ago

#154 - How to predict an images with trained ckpt model?

Issue - State: open - Opened by Amario1306619051 15 days ago

#100 - Doctr meets parseq

Issue - State: closed - Opened by felixdittrich92 over 1 year ago - 8 comments

#100 - Doctr meets parseq

Issue - State: closed - Opened by felixdittrich92 over 1 year ago - 8 comments

#99 - [enhancement] Add beam-search for generating topN decoded results.

Issue - State: closed - Opened by FeiZhuNiU over 1 year ago - 1 comment
Labels: wontfix

#99 - [enhancement] Add beam-search for generating topN decoded results.

Issue - State: closed - Opened by FeiZhuNiU over 1 year ago - 1 comment
Labels: wontfix

#98 - Should self. head be set to len (self. tokenizer), especially when predicting batches, allowing labels of different lengths

Issue - State: closed - Opened by kingqiuol over 1 year ago - 1 comment

#98 - Should self. head be set to len (self. tokenizer), especially when predicting batches, allowing labels of different lengths

Issue - State: closed - Opened by kingqiuol over 1 year ago - 1 comment

#97 - My val_acc is 99+ but my testing accuracy is around 40%. I do not understand what I am doing wrong.

Issue - State: closed - Opened by ceyxasm over 1 year ago - 6 comments

#97 - My val_acc is 99+ but my testing accuracy is around 40%. I do not understand what I am doing wrong.

Issue - State: closed - Opened by ceyxasm over 1 year ago - 6 comments

#96 - NAR decoder error

Issue - State: open - Opened by Gavinic over 1 year ago - 4 comments

#96 - NAR decoder error

Issue - State: open - Opened by Gavinic over 1 year ago - 4 comments

#95 - About charset used in train and test

Issue - State: closed - Opened by wdp-007 over 1 year ago - 2 comments

#95 - About charset used in train and test

Issue - State: closed - Opened by wdp-007 over 1 year ago - 2 comments

#94 - Only recognize numbers

Issue - State: open - Opened by jeeve over 1 year ago - 1 comment

#94 - Only recognize numbers

Issue - State: open - Opened by jeeve over 1 year ago - 1 comment

#93 - Was any sort of image pre-processing done prior to training?

Issue - State: closed - Opened by ceyxasm over 1 year ago - 2 comments

#93 - Was any sort of image pre-processing done prior to training?

Issue - State: closed - Opened by ceyxasm over 1 year ago - 2 comments

#92 - lmdb.Error: data/test/-: No such file or director

Issue - State: closed - Opened by ceyxasm over 1 year ago - 1 comment

#92 - lmdb.Error: data/test/-: No such file or director

Issue - State: closed - Opened by ceyxasm over 1 year ago - 1 comment

#91 - Remove test-time dependency on PyTorch Lightning

Issue - State: closed - Opened by baudm over 1 year ago
Labels: enhancement

#91 - Remove test-time dependency on PyTorch Lightning

Issue - State: closed - Opened by baudm over 1 year ago
Labels: enhancement

#90 - Improve dependency pinning

Issue - State: closed - Opened by baudm over 1 year ago
Labels: enhancement

#90 - Improve dependency pinning

Issue - State: closed - Opened by baudm over 1 year ago
Labels: enhancement

#89 - 8-bit quantitization with PyTorch 2.0

Issue - State: open - Opened by eschaffn over 1 year ago - 3 comments

#89 - 8-bit quantitization with PyTorch 2.0

Issue - State: open - Opened by eschaffn over 1 year ago - 3 comments

#88 - TORCH 2.00 breaks everything

Pull Request - State: closed - Opened by orithecapper over 1 year ago - 1 comment

#88 - TORCH 2.00 breaks everything

Pull Request - State: closed - Opened by orithecapper over 1 year ago - 1 comment

#87 - Loss is turning NaN for PARSeq

Issue - State: closed - Opened by strmojo over 1 year ago - 5 comments

#87 - Loss is turning NaN for PARSeq

Issue - State: closed - Opened by strmojo over 1 year ago - 5 comments

#86 - My GPU is not recognized and the sample size is weird

Issue - State: closed - Opened by cksdlakstp12 over 1 year ago - 2 comments

#86 - My GPU is not recognized and the sample size is weird

Issue - State: closed - Opened by cksdlakstp12 over 1 year ago - 2 comments

#85 - How to reproduce the accuracy of Table17 in the Appendix J？

Issue - State: open - Opened by FaltingsA over 1 year ago - 2 comments

#85 - How to reproduce the accuracy of Table17 in the Appendix J？

Issue - State: open - Opened by FaltingsA over 1 year ago - 2 comments

#84 - Label data format for custom dataset

Issue - State: closed - Opened by cksdlakstp12 over 1 year ago - 2 comments

#84 - Label data format for custom dataset

Issue - State: closed - Opened by cksdlakstp12 over 1 year ago - 2 comments

#83 - Building model for grayscale

Issue - State: closed - Opened by kino0924 over 1 year ago - 3 comments

#83 - Building model for grayscale

Issue - State: closed - Opened by kino0924 over 1 year ago - 3 comments

#82 - Training Suggestions for Cyrillic + English

Issue - State: open - Opened by eschaffn over 1 year ago - 3 comments

#82 - Training Suggestions for Cyrillic + English

Issue - State: open - Opened by eschaffn over 1 year ago - 3 comments

#81 - Training

Issue - State: closed - Opened by manhdoan291 over 1 year ago

#81 - Training

Issue - State: closed - Opened by manhdoan291 over 1 year ago

#80 - Rtd

Pull Request - State: closed - Opened by isaac-kang over 1 year ago

#80 - Rtd

Pull Request - State: closed - Opened by isaac-kang over 1 year ago

#79 - Is there pretrained PARSeq model on the synthetic datasets?

Issue - State: closed - Opened by Xiaomeng-Yang over 1 year ago - 1 comment

#79 - Is there pretrained PARSeq model on the synthetic datasets?

Issue - State: closed - Opened by Xiaomeng-Yang over 1 year ago - 1 comment

#78 - [Question] about training model with another dtype

Issue - State: closed - Opened by phamkhactu over 1 year ago - 1 comment

#78 - [Question] about training model with another dtype

Issue - State: closed - Opened by phamkhactu over 1 year ago - 1 comment

#77 - [Question] about retrain model follow knowledge distillation

Issue - State: closed - Opened by phamkhactu over 1 year ago - 2 comments

#77 - [Question] about retrain model follow knowledge distillation

Issue - State: closed - Opened by phamkhactu over 1 year ago - 2 comments

#76 - Tensorrt gives the same output no matter what the input is

Issue - State: open - Opened by AbhiHegde3004 over 1 year ago - 5 comments

#76 - Tensorrt gives the same output no matter what the input is

Issue - State: open - Opened by AbhiHegde3004 over 1 year ago - 5 comments

#75 - urllib.error.URLError: <urlopen error [Errno 11001] getaddrinfo failed>

Issue - State: closed - Opened by kuldeeps1208 over 1 year ago - 3 comments

#75 - urllib.error.URLError: <urlopen error [Errno 11001] getaddrinfo failed>

Issue - State: closed - Opened by kuldeeps1208 over 1 year ago - 3 comments

#74 - Detection of Multiline Text

Issue - State: closed - Opened by kuldeeps1208 almost 2 years ago - 2 comments

#74 - Detection of Multiline Text

Issue - State: closed - Opened by kuldeeps1208 almost 2 years ago - 2 comments

#73 - Can it be run on machine enabled with CPU

Issue - State: closed - Opened by kuldeeps1208 almost 2 years ago - 1 comment

#73 - Can it be run on machine enabled with CPU

Issue - State: closed - Opened by kuldeeps1208 almost 2 years ago - 1 comment

#72 - Fine tuning dataset size

Issue - State: closed - Opened by AzaelCC almost 2 years ago - 1 comment

#72 - Fine tuning dataset size

Issue - State: closed - Opened by AzaelCC almost 2 years ago - 1 comment

#71 - Unable to get eos_id properly after converting to torchscript

Issue - State: closed - Opened by EMHussain almost 2 years ago - 9 comments

#71 - Unable to get eos_id properly after converting to torchscript

Issue - State: closed - Opened by EMHussain almost 2 years ago - 9 comments

#70 - Unable to reproduce the experimental results in the paper

Issue - State: open - Opened by LittleWhite0208 almost 2 years ago - 6 comments

#70 - Unable to reproduce the experimental results in the paper

Issue - State: open - Opened by LittleWhite0208 almost 2 years ago - 6 comments

#69 - Training model on japanese dataset

Issue - State: open - Opened by Preethse almost 2 years ago - 4 comments

#68 - Package ax-platform

Issue - State: closed - Opened by bmusq almost 2 years ago - 1 comment

#68 - Package ax-platform

Issue - State: closed - Opened by bmusq almost 2 years ago - 1 comment

#67 - val_loss and val_accuracy both increasing with training progress

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

#67 - val_loss and val_accuracy both increasing with training progress

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

#66 - Support for onnx:CumSum to improve Onnx results

Issue - State: closed - Opened by SirMomster almost 2 years ago - 7 comments

#66 - Support for onnx:CumSum to improve Onnx results

Issue - State: closed - Opened by SirMomster almost 2 years ago - 7 comments

#65 - LMDB data sets are the same size

Issue - State: closed - Opened by yiren556 almost 2 years ago - 1 comment

#65 - LMDB data sets are the same size

Issue - State: closed - Opened by yiren556 almost 2 years ago - 1 comment

#64 - Hyperparameters for synth dataset

Issue - State: closed - Opened by Xiaomeng-Yang almost 2 years ago - 1 comment

#64 - Hyperparameters for synth dataset

Issue - State: closed - Opened by Xiaomeng-Yang almost 2 years ago - 1 comment

#63 - Train PARSeq for Vietnamese

Issue - State: closed - Opened by Dive-to-OCR almost 2 years ago - 2 comments

#63 - Train PARSeq for Vietnamese

Issue - State: closed - Opened by Dive-to-OCR almost 2 years ago - 2 comments

#62 - How to create a TRT Model from ONNX or direct from PyTorch to TRT

Issue - State: open - Opened by naveenkumarkr723 almost 2 years ago - 14 comments

#62 - How to create a TRT Model from ONNX or direct from PyTorch to TRT

Issue - State: open - Opened by naveenkumarkr723 almost 2 years ago - 14 comments

#61 - Update Typo in Sample Results ViTSTR Prediction

Pull Request - State: closed - Opened by rakro101 almost 2 years ago

#61 - Update Typo in Sample Results ViTSTR Prediction

Pull Request - State: closed - Opened by rakro101 almost 2 years ago

#60 - Can't fix size mismatch errors

Issue - State: closed - Opened by lnmduc2 almost 2 years ago - 6 comments

#60 - Can't fix size mismatch errors

Issue - State: closed - Opened by lnmduc2 almost 2 years ago - 6 comments

#59 - Question about evaluation protocol

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

#59 - Question about evaluation protocol

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

#58 - Unicode error Portuguese Charset

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

#58 - Unicode error Portuguese Charset

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

#57 - ONNX Inference Decoding

Issue - State: open - Opened by rafaelagrc almost 2 years ago - 3 comments

#57 - ONNX Inference Decoding

Issue - State: open - Opened by rafaelagrc almost 2 years ago - 3 comments

#56 - Sometimes predict long, redundancy and repetitive chars, like 'parseqqqqqqqqqqqq'(ground truth is 'parseq')

Issue - State: open - Opened by WenjunLiu6146 almost 2 years ago - 5 comments

#56 - Sometimes predict long, redundancy and repetitive chars, like 'parseqqqqqqqqqqqq'(ground truth is 'parseq')

Issue - State: open - Opened by WenjunLiu6146 almost 2 years ago - 5 comments

#55 - I had a number of images in a directory and i want to test on the linux what are commands i need for this one

Issue - State: closed - Opened by naveenkumarkr723 almost 2 years ago - 1 comment

#55 - I had a number of images in a directory and i want to test on the linux what are commands i need for this one

Issue - State: closed - Opened by naveenkumarkr723 almost 2 years ago - 1 comment

#54 - Support conversion to CoreML

Issue - State: closed - Opened by cywinski about 2 years ago - 5 comments
Labels: documentation

#54 - Support conversion to CoreML

Issue - State: closed - Opened by cywinski about 2 years ago - 5 comments
Labels: documentation

GitHub / baudm/parseq issues and pull requests