Stefan Falk

Reformer Model for Speech Recognition

5

Coming from tensor2tensor I was wondering whether the Reformer model would be also a candidate for speech recognition? Looking at the examples there is none for ASR. Would it be...

Streaming Transformer Transducer

9

Hi! I am currently working on a streaming Transformer Transducer (T-T) myself (using Tensorflow) but I'm struggling to get started with the actual inference part. I've been referred to your...

Training a LSTM for ASR

2

I intend to try out an LSTM for speech recognition. Looking at the t2t code I noticed that there's a `lstm_asr_v1` hparams-set which is should probably work with a `lstm_seq2seq_attention`?...

I am using the `warp-transducer` successfully on other machines (Ubuntu 18.04) but on one, which is a CentOS, I am getting a Segmentation Fault right at the beginning of the...

rnnt_loss() gives different outputs (increasing) executed on same input

1

I am using https://github.com/HawkAaron/warp-transducer indirectly via https://github.com/noahchalifour/rnnt-speech-recognition and I have noticed something odd when running the `rnnt_loss()` function in eager mode on the same input over and over again. Basically:...

Getting ValueError: Attempt to convert a value (PerReplica ..) when starting training

6

Hi! I am currently trying to start a simple training by following the instructions from the README.md. Everything works up to the point where I want to start the training....

Training is not converging. eval_wer sticks at ~95%.

7

I finally was able to run a training on a single GPU (multi-GPU does not seem to work right now) but the word-error-rate is not dropping. I did not change...

I was wondering under what license the [data](https://github.com/UBC-NLP/aoc_id/tree/master/data) is available. The paper states "Creative Commons Attribution 4.0" but is this also true for the dataset as such?

Multi-GPU gives no speedup for transformer model

8

### Description I am training a `Transformer` model on the `Librispeech` dataset using 4 GPUs with 8 CPU-cores. I have tested the following: #### Single-GPU ```bash export CUDA_VISIBLE_DEVICES=0 t2t-trainer \...