-
I'm trying to transcribe an audio clip which is about 2min long with my Seq2Seq model. The transcription is almost perfect but it stops after a few seconds. How can I decode the entire clip?
I've t…
-
I have some problem with fine tuning wav2letter's pretrained models.
I download seq2seq_tds model baseline_dev-other.bin to a folder.
Running following command:
build/Train fork /w2l/am/baseline…
-
I am trying to achieve some good results on Libri 100 Hour data using transformer + CTC architecture provided in https://github.com/facebookresearch/wav2letter/blob/master/recipes/models/sota/2019/li…
-
## 🐛 Bug
Fbanks computed on same waveform change from one run to another.
**SOLVED**: My bad, I didn't know that the dither parameter, if different from zero, introduces noise. Thank you for yo…
-
I tried preparing data using the script `recipes/librispeech/data/prepare_data.py` and for some reason, it was only writing the .lst files and not the other files.
After spending some time trying …
SY-nc updated
4 years ago
-
## 🐛 Bug
`MelScale` and thus `MelSpectrogram` will not use the correct filterbank when `f_max` is not equal to `sample_rate // 2`.
## To Reproduce
Steps to reproduce the behavior:
The foll…
-
I am running Resnet CTC training using user's audios. I am using a very small number of audios for testing for now. I have got an bug: Could not read file ''
Any ideas about that? Thank y…
-
Hi authors,
I'm using **Librispeech run.sh** recipe. I trained the acoustic model (speech_conv_lstm_librispeech) using 4 GPU 1080ti But I'm facing this error while doing kaldi scoring.
local/score.s…
-
I have a .fil file, from the Parkes FRB010125 data. The file size is a few MB and contains a single burst. Can I use a model of FETCH to find if the filterbank file contains an FRB or not? If yes, how…
-
> The flowchart of our diarization system is provided in Fig. 1. In this system, audio signals are first transformed into frames of width 25ms and step 10ms, and log-mel-filterbank energies of dimensi…