Thomas Delteil comments

Results 33 comments of


                                            Thomas Delteil

problem with Tesla V100

FIX is here: https://github.com/baidu-research/warp-ctc/pull/118

upgrade to cuda9(Volta) GPU arch

can this PR be merged? WarpCTC is failing on V100 otherwise. edit: With this branch I am getting: ``` ./test_gpu Running GPU tests Some or all tests fail ```

how to use get_result?

Can you share the error you get when you use: `document = extractor.get_result(job_id=job_id, api=TextractAPI.ANALYZE)` ?

Re-export edited entities to JSON

Duplicate of #120

Output from MXNet model differs from sample from onnx/models

Do you know how 'bad' this is? in % of accuracy lost compared to the original model let's say on imagenet image classification tasks? Is it due to different implementation...

Output from MXNet model differs from sample from onnx/models

Thanks @spidyDev for the update!

Adding license

I should have this sorted out in the next 2 weeks. Thanks

an issues about your ocr data iteration

@zzdang, this is true, this was an acceptable trade off given the small size of the IAM dataset, and to get the ability to load pre-processed images quickly. If your...

The training results are good, but the test results are very poor.

It looks like you are experiencing overfitting. https://en.wikipedia.org/wiki/Overfitting A few ideas to overcome this and regularization to your model: - you can stop training when the validation loss start going...

Fine tuned model

Thanks @techietrader, fine-tuning your network on a classification task on your own dataset is a good idea. Two things to keep in mind: - Try to create a gold cassette...