Stuck on an issue?

Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.

And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.

Segfault: warp-ctc pytorch bindings

See original GitHub issue

Hello, I’m trying to use CTC for an OCR which uses a similar architecture, following the examples at https://github.com/SeanNaren/warp-ctc/blob/pytorch_bindings/pytorch_binding/README.md.

I was initially having a problem with the types (IntTensor/FloatTensor), which I believe I fixed with some trial and error, but right now I’m stuck at a segfault.

Can you tell me what probs_sizes, and label sizes are? I don’t understand this in deepspeech.pytorch:

 sizes = Variable(input_percentages.mul_(int(seq_length)).int(), requires_grad=False)

Issue Analytics

State:
Created 6 years ago
Comments:10 (2 by maintainers)

Top GitHub Comments

2reactions

ryanlearycommented, Jul 28, 2017

It is CTC loss. See the Graves 2006 paper: http://www.cs.toronto.edu/~graves/icml_2006.pdf
The matrix output from the NN contains the log-probabilities of each character class at each timestamp for each sample in the batch. Are you asking for the probability of the emitted sequence?

0reactions

gentaiscoolcommented, Oct 2, 2018

Is it possible to get the log probability of the all characters after CTC?