mmathew( മിനേഷ് മാത്യു )

@Shreeshrii When I tried your fork to find the wordaccsum its computing the summary/aggregated results for only the first two files even when if supply multiple files as arguments to...

@JasonBoy1 s fix doesn't work for me (I use pytorch CTC, not warp CTC). It is even complicated when you have more than one rnn layer. See this thread [here](https://discuss.pytorch.org/t/multi-layer-rnn-with-dataparallel/4450/9)

Setting `shuffle=False` in the data-loader will cause batches to have same set of samples in every epoch right. I think it might cause over fitting. Ideally we must shuffle data...

Hey if you are planning to train for English word recognition , you can use VGG's synth90k dataset https://www.robots.ox.ac.uk/~vgg/data/text/

see [this](https://github.com/bgshih/crnn/issues/2)

hey, Do you know why the L word global dictionary for SANDY has just 77 words ( 107 -30) ? MOMs classification branch and other baseline models use a much...

``` with PyTessBaseAPI(psm=PSM.AUTO, oem=OEM.LSTM_ONLY, lang="eng") as api: # automatic page seg.; LSTM model ; english language api.SetImageFile(image_name) #image_name is the name of the image file on which you want to...

Has anyone figured how to run table detection alone ?

seems it is his dataset in a pickle file. He mentioned that the dataset he uses is an internal dataset and cant be made public. ( I had to translate...