Igor Vishnevskiy

Results 8 comments of Igor Vishnevskiy

Hello Jerod @weinman. Definitely the (2), the tensor of images packed together as input and run a single GPU-parallel inference on them, producing the prediction sequences (i.e. as a sparse...

Thank you so much Jerod @weinman. I will take over from here. Will check in PR as soon as I get it working.

I also tried to tackle this problem from the different side. Sort of the way it is done for .h5 models generated by Keras. With Keras, the conversion to SavedModel...

@weinman Hello Jerod. Apology for the delay. I was simply busy working on the complex project as work. However I finally got back to it and made conversion to SavedModel...

All that I was trying to do above was all wrong by the way. Estimator() class already has a method: export_saved_model(). However it requires the "serving_input_receiver_fn" to be passed to...

@weinman I just created a pull request. Happy to contribute. Your platform is very nice. Thank you. https://github.com/weinman/cnn_lstm_ctc_ocr/pull/70

Just tried to cut inputs down to 10 images, but left 300 classes. Still issue reproduces. Same low res, same 1 batch per GPU. Issue is definitely caused by the...

Just tried to cut inputs down to 10 images, but left 300 classes. Still issue reproduces. Same low res, same 1 batch per GPU. Issue is definitely caused by the...