E2E-ASR icon indicating copy to clipboard operation
E2E-ASR copied to clipboard

Published 20 hours ago • verified publisher icon HawkAaron

implement the ASR to TensorFlow warp binding

Open zhoudoufu opened this issue 6 years ago • 4 comments

@HawkAaron Hi, I am working on RNN-t training with this E2E-ASR repo, with pytorch binding. and I found you have another repo that warp MXnet and TensorFlow. I suppose that warp provides universal usage for 3 different bindings. While I do not find a plug-in place in the training code. Do I need to change pytorch method manually? What is the right way to implement ASR training with TensorFlow warp binding? Thanks in advance.

zhoudoufu avatar Sep 03 '19 09:09 zhoudoufu

You mean use tensorflow binding in this pytorch training code?

HawkAaron avatar Sep 20 '19 01:09 HawkAaron

I would rewrite the training process in tensorflow, and use the tensorflow binding. I am working on this actually. Just curious, is it possible to do that (use tensorflow binding in this pytorch training code)? @HawkAaron

zhoudoufu avatar Sep 20 '19 08:09 zhoudoufu

Of course you can. Just convert pytorch tensor into tensorflow tensor, and call tensorflow rnnt loss.

HawkAaron avatar Sep 20 '19 08:09 HawkAaron

@zhoudoufu Have a look at this issue. I think you can help.

manish-kumar-garg avatar Dec 09 '19 08:12 manish-kumar-garg